AFAICT, the "Rising layoffs year after year" news lately has been an intentional misreporting of "Layoff rates slowly returning to normal after the hiring spree during COVID stimulation"
But, CEOs figured out that if they blame layoffs on AI, stock go up a lot. Reporters know that anxiety about AI drives the clicks that write the checks.
I do think that AI anxiety is making HR around the world anxious about hiring. That's my best guess for why everyone is finding they need to apply to 500 jobs to get anywhere. So, AI is making it hard for you to find a job not because it took your job, but because HR is reading ragebait and turning it into a self-fulfilling prophecy.
I don't recall there being regular, industry-wide layoffs effecting software engineers for basically the entirety of my career up until the last couple of years. I'm sure my memory is bad and there's data to refute that, but this doesn't feel like "normal" at all.
To non-Americans, "global financial crisis" is the standard name we use for it. Inside the US (at least from my experience, when I lived there a few years ago) it was just financial crisis (without "global"), or the housing crash. Draw from that what conclusions you will about how Americans see their place in the world ;)
I worked straight through the dotcom, ‘08 financial crisis, and covid craziness. But recently haven’t worked in over two years, with nothing on the horizon. Definitely not normal over here.
The GFC wasn't good? It wasn't good for the general economy, sure, but given the specific industry we are talking about that was one of the best times ever. Money was being printed hand over fist building software that did nothing more than emit a fart sound. The opportunities for us with the 'picks and shovels' were endless with everyone trying to strike gold at the App Store gold mine.
It is unclear what you mean. It is true that most software jobs were not writing iPhone apps, if that is what you are trying to say, but those were good jobs. Those who didn't have good jobs for a moment were quickly scooped up by investors trying to create the next big app. It was a feeding frenzy out there for tech workers, despite much of the rest of the economy faltering (agriculture also did very well during the GFC, to be fair).
Maybe you are trying to say that there was some grey beard Atari programmer out there who refused to start writing iPhone apps and couldn't find their dream job banging bits on the old 2600? That is likely, but it is equally likely that they never found that job since either.
Correct. This is also the reason why job postings arent as high as they were during and after COVID. During COVID we had “the great resignation” which meant that people were leaving companies to find better pay after 2020 so companies had to “overhire” to account for the poor retention.
This is what a buddy who works at a major consulting firm told me about the hiring trend, we are returning to all the pre-COVID hiring norms.
This is the same with ANY depreciating asset or currency? Can I wait a year to buy it? Will I get 10x the return by waiting a year? Will my competition have a moat if I can move 10x faster next year? can I save money now?
Idk why you're downvoted. Practically everyone i know is unemployed, underemployed, or failing to find a new job to replace one they don't like/isn't paying enough/has bad prospectus.
Lucky engineer types like me in the last group, but many of my college educated friends in that middle group. Shit is rough
I'll need to look at the underlying data, but it seems that the job "classification/category" should be more important than the raw number. A boost in lower paid/service industry jobs does not mean that there wasn't a loss in a category where AI can be more easily dropped into existing businesses.
Average weekly and hourly earnings were up in May [1], though “real average hourly earnings for all employees decreased 0.5 percent from March to April, seasonally adjusted” [2].
Nominal wages being up rejects the hypothesis that folks are being downsized into lower-paying roles.
It's quite frustrating that they track average rather than median wages. As wealth inequality increases, average will be less and less representative of worker health.
Do you have a link? There's no obvious place to look for this. All I can find is "Median usual weekly earnings of full-time wage and salary workers by sex, quarterly averages, seasonally adjusted". Which is not trivial to extrapolate into a real annual wage. And it's not clear how "quarterly averages" and "median usual weekly earnings" compose....
The median would be interesting than the mean, as you can be hollowing out the middle, leaving more low-wage workers and a few very highly paid ones and the "average" still looks good.
Median weekly earnings of the nation's 121.0 million full-time wage and salary workers were $1,235 in the first quarter of 2026 (not seasonally adjusted), the U.S. Bureau of Labor Statistics reported today. This was 3.4 percent higher than a year earlier, compared with a gain of 2.7 percent in the Consumer Price Index for All Urban Consumers (CPI-U) over the same period.
key point, those 6 figures SWEs earn, they would spend to "local non SWE" economy.
your local datacenter does not care about local chickens or eggs, or private tutor, or pretty much anything at all. not even energy, it is has its own nucleaer reactor nearby. it is one-way economy from now on. you are only a consumer, not a producer. there is virtually nothing you (nor average joe) can provide that "datacenter" needs.
This is - quite literally - the fixed-pie fallacy, which has been thoroughly debunked by now. Read up on basic economics before commenting on this topic:
be careful with "real" hourly earnings. if they're using the CPI then it can be very misleading. One need only compare nominal wages vs nominal rents per square foot for the last 50 years to see that actual wages have dropped by 30% and yet the "real" hourly average wage calculated via the CPI is flat, falsly implying that actual wages have not dropped.
> if they're using the CPI then it can be very misleading. One need only compare nominal wages vs nominal rents per square foot for the last 50 years to see that actual wages have dropped by 30% and yet the "real" hourly average wage calculated via the CPI is flat, falsly implying that actual wages have not dropped.
Yes, if you get to pick prices of arbitrary items to compare against, it's easy to come to whatever conclusion you want[1]. That's why CPI uses a basket of goods, specifically to avoid cherry picking shenanigans.
CPI functions to obscure disproportionate price increases in essentials by lumping them in with cheapening commodities (e.g. electronics) that have a high demand elasticity. It's as much a choice to to use CPI to come to the conclusion things are fine, as it would be picking "arbitrary items" such as the essentials: food, shelter, medical care, and transportation that would paint a picture distinctly less fine.
>CPI functions to obscure disproportionate price increases in essentials by lumping them in with cheapening commodities (e.g. electronics) that have a high demand elasticity.
Right, because it's not "essentials price index" or whatever.
>as it would be picking "arbitrary items" such as the essentials: food, shelter, medical care, and transportation that would paint a picture distinctly less fine.
The BLS makes the constituent price indices available as well as their weights, so you can easily vibecode an "essentials" index.
shelter isn't just a cherry picked item. It's by an far the largest spending category, that and healthcare. I couldn't care less about all the other categories, personally.
>shelter isn't just a cherry picked item. It's by an far the largest spending category, that and healthcare.
The BLS agrees. That's why "shelter" is weighted 35.6% in the CPI basket, by far the biggest item.
> I couldn't care less about all the other categories, personally.
If we're going by what people "care" about (whatever that means), the basket would probably be like 70% gas prices, 20% grocery prices, and 10% for everything else. Empirically speaking, those two are far more salient politically than housing.
yeah, staggering. it's utterly amazing how much people care about gas prices. i can't wrap my mind around it. I mean, would people really rather be out on the street rather than get an EV or god forbid a hybrid?
Gas is misleading - petrol / oil would be better. I don't know if natural gas is going up, but if we mean pump prices of the liquid that goes into cars: that's also used (although with some differences) to heat people's homes - oil heat is extremely common in rural areas. This winter it cost me about 600 dollars in one month to heat a home in the Northeast I kept at 60 degrees except weekends!
Electricity is even worse.
Petroleum also impacts industrial goods prices, flight costs, etc. It's more than just a talking point for drivers.
Gas prices affect very nearly everyone in America. High housing prices benefit the most vocal middle class constituencies (ie boomer homeowners)
Rent is a bigger consideration for most people i know, but most people i know take public transit.
Also keep in mind: people's rents don't suddenly triple in the span of a month when the president starts another illegal war... The shock value of gas, uh, shocks makes it easier to talk about than the slow relentless squeezing of the working class via rent increases that outpace inflation and wages
Much of the western world has healthcare that is free at the point of use and paid for by taxation; one cannot extrapolate consumer inflationary healthcare costs in quite the manner the USA does.
Remember, CPI exists to attempt to figure out how the value of a currency has changed over time. It's not a cost of living index, a housing cost index, a wellness index, or anything of the sort. It is not trying to measure how you are doing in life. It is not there to determine the value of labour, the value of housing, the value of food, or the value of anything other than currency. It is simply compiled to determine the difference in value between $1 yesterday and $1 today, so to speak. As you can see, you cannot use the currency itself to provide that measurement. $1 and $1 are visibly the same; except we know that $1 and $1 aren't the same over time. That is where CPI steps in.
If you want to understand something else, use the measures and data that are focused on that something else. Right tool for the job and all that.
Huh? Housing becoming significantly more valuable than in the past[1] does not imply that wages have dropped.
[1] Mostly thanks to computers/software making housing a more compelling proposition. Historically, people couldn't wait to get outside or over to the pub and would allocate their spending to support that, but now that they prefer to stay in to look at screens, they deem that place where they spend time worthy of greater spending instead.
What would cause an increase in the number of open lower paid and/or service industry jobs while simultaneously reducing the number of openings in tech?
It's crazy that so often I see articles, here on HN and elsewhere, where some pundit claims that there is no AI job crisis, AI isn't replacing any jobs, that layoffs are actually due to post-pandemic ZIRP overhiring, etc.
But then people who work in actual tech companies come in and explicitly say they are not hiring any juniors anymore specifically because AI is good enough to do most of what juniors do, and that senior engineers can now write 3x as much code, etc.
There seems to be a desire for a narrative that AI really just can't replace productive work, and that it's all a mirage. However it seems just like common sense that if an AI can do junior-engineer-level coding work, that a company has less reason to hire a junior engineer.
"But then people who work in actual tech companies come in and explicitly say they are not hiring any juniors anymore specifically because AI is good enough to do most of what juniors do, and that senior engineers can now write 3x as much code, etc."
If you want an anecdote: the media company I work for just started hiring interns and juniors in software career tracks again after a lengthy hiatus.
I have a suspicion that Twitter laying off so much of the software staff was very influential for people with hiring abilities. The company didn't crash, and they (relatively slowly) began shipping new features again. I think that coincided with the pandemic-era overhiring, and we've been working against that combo ever since.
Every now and then, I actively try to make an LLM replace my tasks, and fully do greenfield projects I would accept- I don't see it. It's very good, no doubt. But I have or have been given the project parameters, and just like with a junior, failures in communication inevitably lead to breakdowns in execution.
It does your job, but not completely autonomously so you don't see it?
It requires a lot of guidance, luckily, I mean thank god otherwise we would be goners. The job itself hasn't become easier, but it did change. If you come across failures you update the spec, the guard rails, whatever you use to guide it. It's not a "it produced bad code and now it's forever useless" type of situation.
I hate to be "that guy" but you're crazy if you don't see AI being able to do greenfield projects you'd accept.
I mean, in your defense, last year I think you would have been right, but right now? Codex rocks, as does Claude. I am literally making money shipping a greenfield project to a customer right now. I'm basically a cheap consultant that is incrementally adding features to make something exactly what they want for way cheaper than it would be to do it the old way.
There are hiccups, outages, things to fix etc. but the customer is happy with the output, and the reduced price means they get bespoke solutions rather than some BS one size fits all SaaS app. Then my job is maintenance and effectively "ITSM." Which kind of sucks in some ways, I miss writing real code for real projects, but this is the future going forward. If you want something for your business, you'll generate it rather than pay for it and for now at least, getting beyond localhost requires someone who knows a bit about computers or is willing to learn. Most small businesses aren't willing to learn.
Now, to your point. Is the code all that clean? Nope (and in your defense sometimes I read through the codebase and shudder)... but who cares? Like, for awhile I would go through and frantically edit it, but why? It worked. Not only that, but there's going to be a new model in 3 months or whatever that can clean it up and make it less shitty. I've literally done that a couple times since I started doing this in January.
The customer ain't reading the code. They don't care as long as the the functionality works - that's what counts. The gazillion tests I have keep it stable as I push code, and the CI/CD pipeline removes a ton of the ass pain I'd have without it.
The biggest thing I'm worried about when it comes to clean code and good design is trying to make sure I keep the token count down on these projects so I can actually do meaningful work without burning through a week's worth of tokens in a single day. That, and I like to try to keep a sort of architectural bird's eye view on what's happening...
Like, I'm not sure what niche of the industry you're in, but for the stuff I'm using it for, stuff is working really well with LLMs.
I assume they are working on low stakes software. Does it really matter if you use an LLM to code a scheduling app for a hair salon or veterinary clinic?
Think about all the geneticists that complain about excel re-writing DNA sequences.
It doesn't matter if its high stakes or low stakes. People use the software that generates the results they like. Not the software that is "correct to use".
So to try and understand your position - you are hired by a small? Company in which sector? And you built an app that does what? And I think most importantly - how did you find the gig? Were they explicitly hiring a AI capable person to do X?
I run my own thing since the start of the year. I’m building little tools for an industry I’m highly familiar that needs very specific scheduling software, data tools, tracking tools etc. My old job was a boring as a government bureaucrat.
I started this by doing some work for an old employer that asked me to start by modernizing an excel spreadsheet into an app I made for them like 10 years ago? They kept asking for more though since, and they’re my biggest customer right now. Which is good because I only have the bandwidth for like one more place right now.
I’ve had a few sort of one off things with other places? But I’m working on getting another company in the same industry right now and I’ll be able to adapt most of the code I’ve built here for another company if they end up deciding to use me.
But my biggest value to companies is “I already know this industry extremely well.”
There's no rule that a phenomenon needs 1 cause - the decline could've been junior cuts to start after '22 since they have less seniority and impact, followed by a substitution effect after broad LLM adoption. I don't doubt that remote had an additive effect on productivity via mentoring etc as well.
> There seems to be a desire for a narrative that AI really just can't replace productive work, and that it's all a mirage.
Yes, and juniors aren't known for their productive work in the beginning. That's not their purpose. Their purpose is to do the mundane work, because it is important for them to become less junior and more senior.
That is robbed of them.
Which in 5-10 years means the need for senior developers is gonna shoot through the roof.
> All the software engineers we need for the next 20-30 years are already in the current workforce.
There's no way of knowing what the industry will look like decades from now. Even assuming the prediction that seniors become 10x more productive, that would mean software becomes much cheaper to produce. Does that induce enough demand for additional software that keeps employment levels high? Could be, who knows.
Alternatively, maybe software hits a saturation point where there just isn't as much new ground to cover and employment levels crater. That could happen too.
>But then people who work in actual tech companies come in and explicitly say they are not hiring any juniors anymore specifically because AI is good enough to do most of what juniors do, and that senior engineers can now write 3x as much code, etc.
Yup, that's reflected in the data as well, no need to invoke "vibes" or whatever.
Can confirm. From my experience, since December 2025, state-of-the-art models can replace a lot of need for junior, medior or senior developers.
This meant that, in my example and given the same scope, one junior person with AI could easily handle multiple projects in 2026 that would require 2 or more medior people as recently as 2025. The only reasons for involving multiple people were time, learning opportunity, or increased scope - not skill.
If we paused AI progress today, and ignored the option of adding agents recursively, the current state of things still provides several alternatives to any honest & informed manager:
- Altruistic or optimistic: Keep all current jobs, and move the scope until there's a need and/or funds to hire more devs;
- Pragmatic or neutral: Keep the current jobs, but not hire anyone else as long as the scope is balanced against time (our current situation);
- Cold or pessimistic: Lay off anyone who doesn't fit the leanest model at current (or acceptably lower/higher) scope.
I don't know whether this translates into a "job's crisis", and I can only hope it won't. But hope is not a career strategy, scope is not inexhaustible, and the political, economic and social pressures are quite strong even if you removed AI.
I think that's coming at it backward. The article lacks data. Nobody should be expected to run around writing the article that's not there. As it stands, it's unsubstantiated junk.
There's no report. This "article" was three sentences and a chart that doesn't even break down crude occupational categories, let alone anything more specific.
> However it seems just like common sense that if an AI can do junior-engineer-level coding work, that a company has less reason to hire a junior engineer.
I mean, if we want to not talk about economics that’s fine, but can the AI actually do junior work at the same price? What if we don’t look only at quarterly reports, and instead include the value of having people knowing about the business having to explain it to others, who then learn it and can improve it over time?
I think it’s clear the AI is strong, there’s no doubt about that, but that’s not the whole picture.
This is a collective action problem. Its too easy to claim the answer is "we'll hire our competitors juniors when they become seniors" and then every company wants to do that and not train their own juniors. Soon no one is willing to train juniors because theyll just leave for the companies that dont train juniors.
The point is not whether it’s a mistake to not hire juniors, but whether it’s actually factoring into the hiring/layoff decisions at tech companies. Many claims are being made that no companies are actually changing hiring/layoff decisions on account of AI, and are using it as a distraction to avoid admitting their mistakes over hiring. That may be true, but many managers and execs actually do seem to consider AI a sufficient replacement for much of their engineering team, and are stalling hiring/prompting layoffs because of it.
It's not so much a question of whether AI is strong or not. It's a question of whether the tradeoffs (theft of intellectual property, coal burning, lack of transparency, stealing water, rising energy prices, global surveillance, etc.) are worth the outcomes. It's not even a serious question.
If AI was truly strong, we would be seeing signs in the job market. And we would certainly be seeing a lot more subscriptions and demand for these services. For most people, AI does not improve their lives. For a lot of them, especially younger people, it makes their lives much harder and sadder.
> If AI was truly strong, we would be seeing signs in the job market.
Recalling past business cycles, the job market can look "strong" with low unemployment, even while a lot of creative destruction is going on. Look at the 2004-2009 cycle, where the economy looked strong until some time in 2008 even though a lot of money was being wasted on bad projects.
It's only when the tide goes out that you see who's swimming without trunks. So unemployment doesn't typically rise until the recession part of the cycle.
> include the value of having people knowing about the business having to explain it to others, who then learn it and can improve it over time?
It's always been difficult to put a number on that value, which is the problem for the MBAs running the show. There's no number on the P&L assigned to tribal knowledge, and improvements that can be made by those with that knowledge and experience within the business.
It's a mistake that businesses keep repeating, over and over again, yet never actually learning the lesson. And now the industry is going to repeat it again, until there's enough pain that they realize that lost value and start rehiring again.
In my experience, AI raises the ceiling but also lowers the floor. It can make experienced people more productive: they can vet the output. The juniors? Not so much. I've worked with guys who write prompts that are literally "fix it." The result is about as good as you expect.
"Productive junior" is vastly different than a productive software developer with a couple decades of experience. What they produce will differ in quality significantly, AI or no AI.
well said. my wife's company has been having layoffs particularly from the overhiring during COVID. They went from 10 billion down to 1 billion thus far.
And now it becomes so good, that you want to have it. Which means companies have to deal with the budget increase which they will recupe.
How? By making a 5 person team a 4 person team + AI.
If i think about my co-workers (not excluding myself) from last 15 years, there was always someone you would accept just because it was better than not having that one person. If i can now replace them with more tokens/better models, man i wouldn't hesitate (of course i know what this means on a person to person level :/)
The AI job crisis is large tech employers firing workers to free up cash for AI capex, and those workers competing with everyone else in a competitive job market.
I’ve been unemployed and actively looking for a job for about 6 months, the longest stretch of active job searching I’ve ever done in my career. Several close friends who work in tech or tech-adjacent fields are in the same boat. Anecdata on Hacker News or LinkedIn tells the same story.
A chart showing “total number of jobs” is not meaningful. I took a temp job in a metal shop to make ends meet while I wait for the endless rounds of interviews I’ve now gone through with 4 companies. It pays less than half of what I was making before, and I am barely making ends meet. It’s not sustainable, even though the pay is more than fair for the work.
There are also a lot of job openings for home health care workers, or seasonal resort workers, that used to be filled by immigrants. Those jobs are not going to be taken by any of the engineers who just got laid off by Meta.
I have the strong impression that people who write articles like this are very disconnected from the reality of the economy right now, and that their curiosity ends at the line on the chart they cooked up to make a contrarian point.
I am very sympathetic to your situation, and fear something similar happening to me. But the article is addressed to investors, who are interested in aggregate labor demand declines.
"A chart showing “total number of jobs” is not meaningful."
It is meaningful in answering the question being asked, which is whether the hype around labor displacement, which has been growing for nearly five years now, is actually occurring in a way that would justify some of the higher valuations for AI firms.
The "AI jobs crisis" is generally understood to mean a sustained downturn in demand for all labor due to AI substituting for labor across a huge range of tasks.
Same boat, used to do backend infrastructure but have now been doing basic IT support at a non-tech industry; on paper I'm still employed, but between making not enough to make ends meet anymore and not even being fully engaged with the work, I feel effectively unemployed.
That's true! In 12 months, I got two contracts: a 3-month fixed-term contract and a previous 10-day contract role. Let's say that the landlord decides to sell the place I'm renting, I won't be able to rent another place due to unemployment.
I run a job search site and I don't see a crisis in terms of job openings even for SWE - but there is very clear signal that AI is deeply getting embedded in every SWE job.
Thanks. Yes, entry level roles appear to be under pressure, and middle-management roles in the Bay Area may be seeing similar pressure. (anecdotally) What is less clear is whether these are early warning signs of a broader shift, or simply cost cutting measures to account for AI spend.
IMO, AI coding tools have materially improved SWE productivity compared with two years ago. The open question is whether demand for new products and services will expand enough to absorb that productivity gain. If demand does not increase correspondingly, then it may only be a matter of time before we see meaningful job market pressure, starting with SWE.
In the news I hear about tech layoffs, but I am having a difficult time trying to hire high quality senior SWE's. Perhaps it's not the senior's that were hit by the layoffs, or I'm just looking in the wrong places.
The big job sites feel rife with unqualified applicants and outright fraud, leaving me frankly unsure how to proceed.
It feels like a crisis of candidates-to-job-openings matching, and AI has made it much worse.
Just a minor suggestion if applicable: My site didn't pick up your listing - Corvi has a direct integration with Ashby. I recommend enabling all of job boards in integrations to get your posting to job search engines beyond popular ones.
Oh, I'm on the job market, unemployed and probably wont apply becasue of the contract. That's probably your reason for not getting the quality of candidate you're hoping for. that hourly range goes pretty low and doesn't give confidence that a potential employee won't get low balled
I am a manager with former FAANG experience followed by mid-size comanies. Most people here saying there is no AI job crisis dont know what they are talking about and talking out of their ass.
I know plenty of people looking for job for months at this point. I am looking for 3+ months and have a few leads finally but nowhere close to offer. Remember, people have to live in these high cost GEOs (SF, Seattle, NYC and others). Since COVID, the US economy is effy at best with housing prices doubled, rents and household expenses doubled. Just for something simple it costs $200 subscription like mowing the lawn. Pulling stats out of some dashboard that a rando made is not how people live day to day life. Remember, people lie with statistics all the time.
So yes, there is a significant job crisis. Because Ycombinator invested in AI companies this narrative is suppressed here and moderator DANG (one of the worst) heavily moderates anything critical of AI and YC.
Personally, this makes me suspicious of calling out mods for censorship. This, in turn, makes me suspicious about the truthiness of any of the information presented.
I'm not agreeing or disagreeing with any of the opinions or experiences expressed.
Yea, we're basically not hiring anyone that isn't a senior developer already. That's going to be a huge problem eventually but not my problem to deal with.
My best advice for folks that want to get into software now is be willing to do it cheap for awhile and then jump once you've developed some skills. If you were getting into this industry for the money you're properly fucked and I hope you didn't load up on debt. If you're passionate about building stuff there's still room but the path forward is a lot murkier.
The reality is most firms are running out of projects to take that make economic sense.
Note: ECONOMIC SENSE. This has nothing to do with refactoring for the sake of refactoring. Its all to do with earnings growth with respect to the cost of capital.
Me and most of my friends from school started in low paying consultant jobs, in fact most from my class dont do software at all only 5 of us had jobs after education ended, this is like 12 years ago, it was also hard back then to get a job, maybe people forgot or something.
I have been out of the tech startup labor market for many years. But when I was doing hiring in a startup around 2017, we had to put specific, conscious effort into accepting inexperienced candidates. The default mode for hiring is finding like-minded individuals whose experience overlaps with the existing team.
Using AI as a scapegoat for hiring freezes or layoffs is just more FUD. Like all propaganda campaigns, the purpose of fatigue is to induce an amnesia effect.
This has been my experience as well. I tried my hand at switching jobs in 2018 and it was a tough market. It feels like people don't have any concept of transferable skills, and just want a clone of someone who just left and can pick right up from where the other person left off, with no on-boarding period. Nothing seems to have really changed in the intervening 8 years.
I can make more money doing HVAC but I'm tired of being on hot roofs.
> the path forward is a lot murkier.
If you're just here for the money go somewhere else. If you're here because you love computer science then ignore these people and do the work. If you can't find a company get a dayjob and do it for yourself.
> If you're passionate about building stuff there's still room but the path forward is a lot murkier.
Definitely feel the murkiness. I've been programming as a hobby for over ten years and only recently started wanting to do it professionally. I'm actually wondering if there's a path for me.
What people don’t get that traditional software jobs are gone.
There is no need for developers, testers, PMs, TPMs, devops, leads anymore. Communication burden and structure imposed by these roles is too high when you have AI. It doesn’t make sense to tell somebody to do something and for them to type it into the AI. You can type it into the AI yourself without wasting the time.
There is new job - software creator. Think of it as a single person replacing a team. This job requires different mix of skills, and different level of autonomy. Hiring needs to be adapted to this really, and people need to adapt. Some will shine in this new world, and those who still narrowly think of themselves as specialists in the specific old role are going to be jobless.
same applies for seniors as well. ther isn't much distinction of senior vs junior human dev (as in cost and efficiency) compared to AI-dev (cost and efficiency). more so, at current imrpovement rate. in couple more years you would not need seniors anymore either.
As long as architectural decisions have long-term costs and you need human taste and judgement to speculate on what business needs will come about in the next few years you'll still need human engineers.
What exactly was your company doing such that the concurrent development of features and infrastructure by 30 people is now done by 2? Each engineer is monitoring 15 Claude swarms?
This seems pretty unlikely. If it turns out to be true then you don't need a junior or senior dev you can just get a random person from the street and they could do the job.
tbh, it was always kind of difficult to get a job as a Junior engineer, I had to work for free for almost a year to get the job then slowly grind my way up for higher salary
It's really funny that the first thing in the article is a graph showing a long downward trend with a tiny, brief uptick circled at the end and then the entire article appears to be written about that.
I don't see an uptick. Demand has been steady for the past two years after falling from the post-pandemic frenzy. That fits with my personal experience of the tech industry over the past several years.
People are right to point out that hiring is nothing like the post-pandemic years, but it's not clear that it's any tougher out there than, say, 2018. This is from the perspective of someone with a lot of industry experience, though. I can't speak directly to the experience of junior engineers.
Bold claim to say "no signs" based on non-contextual numbers. I think I recall somewhere that most jobs added were in healthcare.
>> The May jobs report reinforced this with nonfarm payrolls jumping by 172,000, confirming that there are no signs of workers being replaced by ChatGPT.
It's not just that. It's anyone who wants to live longer and look younger. Pretty much anyone with the income to afford it. As society becomes wealthier , it means more $ spend on elective procedures and healthcare overall. Wellness clinics are a huge deal now.
Yeah I'm starting to think the BLS needs to do "Boomer aging adjustments" to jobs numbers, in a similar manner to how they do "seasonal farming adjustments". Until we get through the Boomer population bulge, healthcare is going to keep adding jobs for quite a while regardless of how the rest of the economy does, but that doesn't necessarily mean the overall labor market is healthy.
If you leave out healthcare, 2025 had massive job losses overall, with Boomer bedpan cleaning bringing the net number up to just above zero.
If there is an increased demand in healthcare jobs that will increase wages in healthcare which will pull people out of other jobs and into healthcare in a healthy labor market. I’m not saying whether or not the labor market is healthy, but this adjustment wouldn’t help you figure that out.
Also Gen X isn’t that much smaller than the boomers, and millennials are the largest generation ever. Plus all generations aster the baby boomers have fewer children per couple to take care of them, so demand for healthcare jobs isn’t going to drop anytime soon.
> the BLS needs to do "Boomer aging adjustments" to jobs numbers, in a similar manner to how they do "seasonal farming adjustments"
You’re comparing a low-frequency trend with a high-frequency cycle. The latter has lots of data to characterize it. The former may be secular or may be a slow cycle; nobody should be adjusting for it in the base data.
There are upward and downward revisions this year at least, it seems like a mixed bag with many policies juicing the economy. [1]
Job opportunities differ by state and de-growth hostility to business policies and crony investments. Where I am, layoffs and offshoring continues. I hear new grads are increasingly opting for the skilled trades, which is interesting given they aren't getting use out of their degrees.
> There are upward and downward revisions this year at least, it seems like a mixed bag with many policies juicing the economy
Since Jan 2025, the only two upward revisions have been in the last 3 months. So we’re both correct here. This year (3 months of data), has been a mixed bag. But on a longer time horizon (last 6, 9, 12 months) the revisions have mostly been negative.
Yes, I hope I didn't contradict you, the economy has been shite. Was just trying to leave a positive note on recent developments that will continue, God willing.
Well... if you think in terms of a society spending its people on doing various things, spending more people on healthcare could be a good thing. It means we're getting food grown and stuff made with fewer people, so we can spend more people on making sure that everybody lives, and lives healthier.
Healthcare is a huge sink if you have the income to afford it . People will spend hundreds of thousands of dollars on various protocols to look younger, beat aging, and so on.
Workers taking on new/different roles isn't the same as being replaced. Workers have been taking on new/different roles since at least the advent of agriculture, so that's nothing new. Being replaced would be something new, but the data doesn't support it.
How much time has to pass before we can dismiss the widely cited predictions by the media (and also many popular bloggers and podcasters) that AI will lead to mass joblessness? Those people need to at the very least acknowledge that their forecasts are so far wrong.
To answer your question: It depends on your claim. All claims of mass unemployment thus far are evidently not true. All claims of mass unemployment in the near future: We'll have to wait and see.
The prediction that AI will lead to mass unemployment is largely a widespread fear more than an observation.
AI washing: On the one hand, you have periodic mass layoffs that could be related, but also could be "AI washing", with the companies citing AI to sound progressive when, in fact, they just over-hired.
Hum... That's a huge conclusion from a single graph, I expected to see more data.
The jobs openings / unemployed close but a bit larger than 1 is the expected value when nothing is turning the economy completely upside down right now. It means there is no new shock that is too fast for people to adapt.
What is surprising, yes, because I could easily name half a dozen shocks that I would expect to see there. But it seems US people are adapting quickly right now, or things are canceling out.
First of all the May jobs report was mostly in temporary workers possibly due to the World Cup. Second as already noted the jobs are mostly in healthcare. Third job openings does not equal employment. These numbers have been diverging for a while likely due to people holding multiple jobs. Also I believe the evidence suggests the job crisis is due to WFH and not AI https://news.ycombinator.com/item?id=48326721
BLS doesn't look at job ads when compiling "job opening" data. Their method isn't perfect (nothing in life is), but far more comprehensive than you give it credit for.
The results of them actually talking to businesses and asking questions that are more than "did you have a job ad posted?" You are hardly the first person to imagine that job ads aren't representative of actual job opportunities. Obviously they are going to put in effort to avoid those weak signals.
The jobs AI crisis is me being stressed and overworked as management demands more AI, showy (but shallow) "look I made a skill" presentations grab attention, and my PR review queue grows every day as people generate more code than ever.
But to keep this out of a low-value vent, my experience has been that the _threat_ of it is there, but in my small corner of the world/industry, lots of layoffs that would have happened anywhere may just be categorized as "AI" layoffs, but the wild manpower reducing benefits aren't really there. The larger an org gets, the more of your job is dedicated to human stuff, and you can just get some of the code part done a little more quickly.
Would be interesting if we could measure how much effort is put into agentic coding harnesses, frameworks, and theory, vs labor saved using them.
Have never worked in my life like last year, and there's no signs of slowing down.
At least I don't have to review PRs, we have built a team around the concept of trust and not needing to check each other's work.
There's simply too much work to start nit picking in PRs, everybody's very talented and a good engineer you can trust. They ask for feedback when they want it.
And no, we're not producing slop, if anything AI has helped improve the codebase a lot and the harnessing has definitely raised the bar at all levels.
But sure as hell, the job sucks now, and I hate it. And it's not just due to the amount of work.
We have built our team around not needing team visibility. We're 5 developers working on 7 products used by 5 different companies.
I *don't* want to have to know what others are doing nor how. This may work on teams that work on a narrower scope and lower rhythms, cannot work for us.
I'm a consultant, and had my first conversation about an AI clean-up job this week. I'm also just starting another gig analysing LLM output, my sell is that the analysis is hand-coded, as they weren't able to do it themselves with LLM support.
On the other hand, I'm just finishing an agent-heavy piece. After getting it set up, it's been some of the most mindless and soul-destroying work I've had the displeasure of in a while. This stuff will be near minimum wage in a few years, totally unskilled babysitting.
AI really hasn't been all that bad for work, by volume at least. I know where I want to focus my efforts though.
I operate in a bit of a niche, in terms of my clients. They're all typically people I've worked for full time in the past, or closely related. There are some fundamentals of that business I worry about, but I'm able to do things they can't, and I'm looking alright for the year. That's as good as I can generally wish for.
Longer term, I don't know. I'd happily take something more secure if it came along, as long as it's not childminding for an agent. Super busy and bored out my mind the last few weeks, the worst sort of work.
The expectation of linear presentation of change in a bistable system is gobsmacking here.
If this kind of argument were generally valid, it would imply that:
- all change neither accelerates nor decelerates, which is absurd, on the face of it;
- the initial stages of a deep change are always surface-visible; for instance, cancers announce themselves when they begin to gestate, rather than when they metastasize
- A few recent points of data of questionable significance outweighs a hypothesis with considerable support from reason, intuition, and other (unpresented) data. For example, the plight of recent CS grads, which _is_ new, and _is_ on graphs, just not the one the author here chose.
So, since these implied claims are self-evidently _false_, it means that the author would, at a minimum, need to provide an explanation as to _why in this one instance, these considerations do not matter_; for example, the author could have argued that the graph positioned at the center of their argument is the one to look at (as opposed to, say, recent CS grads,) but that _itself requires further argumentation._
It also does not account for the other obvious possibilities; e.g.,that there is a delay between the (as it were) lightning and its thunder; or that even strongly nonlinear effects would have shown up by now in the metric chosen; etc. But since these contributions were not included in the original post, I have no choice but to discount it.
Well, I'm on a small two person team.. for the last few years I would have loved to have 2-3 people under me.. I've been underwater constantly, but we couldn't afford it.
Now I have no need for anyone from a coding perspective.. I can keep up with multiple clients requests with a breeze. I don't have to manage anyone. I type of my phone while I'm on a walk and work gets done for me.
To me, it doesn’t make a lot of sense to hire fewer juniors because of AI. This assumes that the only thing AI can’t replicate is experience. This in itself already seems questionable, but most importantly what about creativity, relating to humans, common sense, etc.?
If you had to hire a concierge for a hotel, would you rather take a guest-oriented, quick-witted junior or an AI? If you could either take a Waymo or an 80-year old driver, what would you take?
I don’t believe hiring managers think that one-dimensionally. Roles are unique and in some roles experience is more important than in others. Plus, juniors already balance their lack of experience with lower salary expectations.
The much easier explanation for the anecdotes: companies are more cautious at the moment and if you only have a few positions to hand out, you rather take the proven hand than risking it on someone who hasn’t shown yet that they can do it
For US SWE labor, off-shoring was and still is a big contributor. I do think AI is a factor too though. You almost get laughed at asking for headcount now. For junior positions in particular, there are close to zero openings.
AI has grown dramatically in capability since last year as well so I'm not sure 2025 data holds today.
It was difficult getting junior headcount in 2023 as well.
That said, I agree with you that AI capabilities have grown massively, but the only way you aren't burning tokens with marginal RoI is having someone with credible domain experience using Claude Code or Codex hand-in-hand.
Basically, no one wants to spend $200K TC on training a junior who will most likely be using an AI tool already. The math works out more positively at a significantly lower TC.
These analyses always crack me up. "There's 1 at least 1 job per person, why are people complaining" is the vibe I always get.
As though the decades of work I've put into my career means I should want a job as the hamburger man or working as a ditch digger. No shade to burger flippers or ditch diggers, but these are not jobs that I'm trained for, nor are they jobs that I remotely want to do. So for me, aviation expert, programmer, ML engineer, weird IT generalist, guy with a math degree who speaks a couple languages, and isn't exactly fully capable anymore, the idea that jobs in some other field (like an RN or an Oncologist or Electrician) are just something I can pivot into is just hilarious. It's such a shallow and ignorant take. Even the premise that all jobs are equally distributed and there aren't jobs that are more or less in demand at any time is a real funny take too.
I'm so glad I started working for myself, because honestly, seeing this dogshit analyses from supposed experts means I'll be able to keep making money for a long time just be actually trying when I need to think about something.
Gedankenexperiment: If it turns out that the uptick in hiring is due to having to clean up after the first generation of vibecoded apps, is that really a net gain?
If America becomes wealthier due to AI, it will mean more follow-on jobs such as people using their newfound fortunes to remodel their homes, consume more , vacation, etc. Companies will expand and hire. All of this creates jobs even if AI may also destroy some jobs. The net result is more jobs. There is a huge market for upper-middle-class people in their 30-50s to look younger. This means more health clinic jobs and demand for pharma.
Theoretically, that's true. In practice the gains were limited to the wealthy who use that money to fund increasingly deranged start-ups* until the market crashes. Then rinse and repeat. The failed start-ups effectively waste any increased productivity leaving everyone about where they were before.
*They fund these start-ups to get a good return on investment so they can get even more money. As the economy overheats the number of places to invest with a reasonable return falls so they are left with the high-risk stuff to invest in. I'm not sure what they want the money for, though, since they could already afford most of the things I would find useful...
I think funding deranged startups is a type of consumption, and it does get money back into the economy the same way funding a remodel does. Maybe the reasoning to do so is different, and maybe the deranged startups add more capital than a bathroom remodel, but then again, bathroom remodels probably tank less often than startups, and worst case you can always read a magazine in your new bathroom. hard to do that in a dead startup
Sort of, that it crosses so many lines makes it seem like it must be 6X, but it peaks at 230 based on a baseline of 100, so just 2.3X their baseline. Still a ton, but not as much as I thought at first glance.
The "edit graph" button reveals some pretty sweet ways to mess with the chart, yet I don't see an option to fix the Y axis at zero. Weird.
Maybe words work better?
There were roughly 28 months' worth of tech job postings within the 15 month period from July 2021 to October 2022.
If you change the baseline to a rough average of the last 2 years, 0.66, that ratio becomes 42 months' worth of job postings within 15 months.
I'd be curious to see what the numbers look like as a percentage of the existing tech workforce. Like, if there are 100 workers and the number of job postings doubled from 5 to 10, that's a huge deal. But if there are 1000000 workers and the number of job postings doubled from 1 to 2, well that's not a huge deal.
Pretty shallow analysis. Doesn't account or what sectors the new jobs are occurring or account for the extended length of unemployment many are experiencing or underemployment. If only he had asked chat GPT, it would have provided more context and nuance than a single chart.
I wonder when we’ll reach diminishing returns on new AI models. Haven’t tried Mythos or Fable yet but it seems like Anthropic is already priming itself for this by calling for a “slowdown” of AI development.
I think it's less of a technology problem and more of a money and investment problem. They called for the "slowdown" right when they filed for IPO. This allows them to both triage the money bleed while under the microscope as well as make their competitors look unethical and dangerous if they don't heed the warning to slow down before we lose control of A.I.
At least in cybersecurity, we already have seen diminishing returns with newer models.
At this point the harness/applayer matters more, as different models perform better or worse on exploit classes depending on the prompt, tuning, and various other parameters.
Of course, by the time HN hyperfixates on a topic, it's already been executed on and HN is too late.
Yea, in many usecases the tooling space is increasingly sophisticated context management such as fine tuning domain specific mappings into the model so that it is able to work directly with a compressed form of some data without needing to decompress into the context.
In larger models, these fine tuning techniques work more reliably/robustly. Because of this many usecases tend to prefer larger models. It is possible to work the same behaviour into the smaller model, but it requires more effort. But it's one-time. And smaller models are usually much cheaper. People make a tradeoff along this curve.
This is observed at few-B scale upto hundred-B scale. No way for us non-anthropic/openai to fine tune beyond that of course.
You can't draw any conclusions based on "job openings" without dealing with, or at least addressing, "ghost job listings". There are several issues here:
1. AI ATS systems have made posting jobs "cheap", such that too companies post jobs that don't exist (ie "ghost jobs") to keep up appearances they're hiring or just to keep people in the pipeline in case they hire. This is a huge waste of everybody's time and should be illegal;
2. The hiring process itself gets increasingly Kafkaesque. AI screening, automated online tests, unpaid take-home work, etc. You have to get pretty far until a human gets involved. 10+ years ago this didn't happen because people needed to be involved much sooner and that's expensive;
3. In a lot of companies, getting employees to interview people is unpaid extra work effectively. They say it's important. You might even get dinged for not doing it. But anyone who has done it realizes pretty quickly a bunch of people who shouldn't get interviewed are getting interviewed and management doesn't care, even though employee time is expensive, because you essentially have to "make up the time" so it's still "free";
4. Even if you go through all that and get hired, you get laid off within a year such that income isn't dependable and you end up wasting a ton of time on the job-seeking process itself.
I've been thinking about this recently and high-information is part of the problem. In years long gone, it was hard to reach applicants so you'd have a small pool of higher-relevance candidates applying for a job. Say 10 people applying for 10 jobs. The odds were better. It was less work on everybody's side.
But now you have 200 people applying for 200 positions. This wastes everybody's time but the problem is that companies have offset this by pushing filtering onto these automated systems. People still need to enter all their bio information, etc. So it's just much more inefficient inherently even if the job opening is legitimate.
The unemployment rate is cooked. It doesn't capture underemployment, people who want full-time employment and don't have it and people who don't make a living wage.
Long gone are the days when the vast majority of people just have one job. Now it's 2-3 part-time jobs because companies are exploiting a legal loophole where they only have to pay benefits for full-time employees. And then you have people doing "gig" work, which is often sub-minimum wage when you factor in expenses (eg driving for Uber and not factoring in car wear).
On top of all this we, across the Western world there's an increasing youth unemployment crisis. In 2008, entry-level jobs basically disappeared overnight and never came back. Well, that just got worse post-pandemic.
No, you're wrong and have no data to back it up. I, on the other hand, bring valuable personal experience that says AIs are useless. I asked it to do X and it did X, but it did it differently than I would do it. This means it is fundamentally flawed and complete utter trash. If it can't do X exactly how I like it without any guidance whatsoever it is not ready for prime time and never will be because it lacks "intelligence".
If every company only hires seniors from now on, and if we assume that the number of senior level jobs is steady, why do I not perceive an increase in leverage like during covid?
The tech co I work for just massively scaled token usage after the copilot price changes. If there was a chance for ai to replace anyone, it's gone now.
Maybe AI will finally be the tool that allows us to get rid of some of the people we have who do nothing more than push paper around. Maybe. But somehow I doubt it, at least not in a typical big corporate environment. And I have zero concern about us letting actual software devs go. Things will have to change pretty dramatically before we get that far.
You mean the middle management? I have been in environments where they were almost literally made up of pencil pushers. Wouldn't be too sad to see them go. Only half joking, but it is written in jest.
Nope, just people who basically swivel chair information from one place to another. Useful in some way, should have been automated a long time ago, and yet persist.
But yes, middle management would qualify ;-). My manager seems spooked by LLMs. Loves to use them to write his emails, but seems to internalize that since they're doing his job for him at this point, his boss may figure it out.
I used to think comparisons of AI with the web where ridiculous, but increasingly it looks like they're not that dissimilar as far as how they change how we work. But as someone who graduated college during the bust there was a loooong gap between peak hype and things like online banking and e-commerce becoming standard.
Even if AI is an absolutely bubble, and SpaceX, Anthropic and OpenAI all cease to exist in a year... there's simply no way that AI has not fundamentally changed work. Even if I was forever pinned to the local models I'm running and the agent harnesses they use, I would never write code for work the same way.
But I lived through the rise of the web. I remember serving dynamic websites through cgi (which meant a new instance of an interpreter was spawned per user session). I vividly remember great JavaScript books saying things like "never use JavaScript for core functionality". I recall Java engineers saying Ruby on Rails was a toy and would never take off, that Python offered nothing over Perl and that "rich web applications" where never going to replace app native interfaces. I remember when the MVC pattern from the early Smalltalk days being dusted off and repurposed for web applications, completely changing how we designed software for the web.
And all of that is just software. It wasn't until the pandemic that ebooks replaced print books in share of academic library circulation (reversing a decades long trend of reduced circulation).
In my daily use of agents for coding and other forms of problem solving, while it is a wild accelerant, it's also clear we have not even started scratching the surface of how to think about building things with these tools.
I suspect we'll adapt to AI faster, but having lived through one major tech revolution, transforming work still takes some time. I'm not surprised we don't see an immediate jobs crisis.
Not to mention the completely separate topic that huge classes of employees were not and are not all that productive, so boosts in productivity don't imply lost jobs. That would require a boost in productivity combined with pressure to create concrete value with less, looking at the SpaceX IPO we're still a ways away from working about how efficiently we create concrete value.
Companies learned one weird trick, that's why you don't see job opening collapse. Most of the job openings listed are fake. Companies aren't hiring despite showing signals that they are; you can ask any engineer that's been laid off over the past year and they'll tell you the same story of how bad the job market is right now.
the AI job crisis will be because the AI bubble will pop and th economic fallout is going to be tremendous. dotcom and GFC might not be as bad as what the AI bubble popping might be.
maybe look into industries AI is best at automating? like constant layoffs in Software? 160,000 people lost jobs in just 2026?
the claim "hey there is no AI job crisis", when previous SWE of 6-figures now takes job dishwashing in McDonalds + one more gig as Uber driver + food delivery gig is "job creation! now they have 3 jobs!". does not make any sense.
Even if they're more productive, the amount of work to be done has also increased significantly. Not only is there so much work to be done on AI integrations but all of the other products have bug trackers exploding with new privacy and security vulnerabilities that need to be addressed.
Some of the security people I work with claim that without the productivity gains from AI, there is no way they could keep up with the bugs.
causation can only be expressed within your analytical framework.
if it not good, then whole world is a chaos and nothing causes anything.
causation is pure interpretation.
correlation is there regadless of what you think of it.
correlation is interpretation-free.
point being, simplistic correlation != caustion is a mid-wit position.
AFAICT, the "Rising layoffs year after year" news lately has been an intentional misreporting of "Layoff rates slowly returning to normal after the hiring spree during COVID stimulation"
https://blog.glassdoor.com/site-us/wp-content/uploads/sites/...
https://www.glassdoor.com/blog/worklife-trends-2026/
But, CEOs figured out that if they blame layoffs on AI, stock go up a lot. Reporters know that anxiety about AI drives the clicks that write the checks.
I do think that AI anxiety is making HR around the world anxious about hiring. That's my best guess for why everyone is finding they need to apply to 500 jobs to get anywhere. So, AI is making it hard for you to find a job not because it took your job, but because HR is reading ragebait and turning it into a self-fulfilling prophecy.
I don't recall there being regular, industry-wide layoffs effecting software engineers for basically the entirety of my career up until the last couple of years. I'm sure my memory is bad and there's data to refute that, but this doesn't feel like "normal" at all.
If your career was entirely post GFC you lived through the good times. The GFC and dot com crash were not good.
So after 2007 for people like me who arent used to the GFC acronym. Which in my case, I was 17 in 2007 and “shielded” from all of that.
For others like me: GFC = Global Financial Crisis
I went through the dotcom crisis and never heard of GFC until today. I've always seen it as the finsys crash.
To non-Americans, "global financial crisis" is the standard name we use for it. Inside the US (at least from my experience, when I lived there a few years ago) it was just financial crisis (without "global"), or the housing crash. Draw from that what conclusions you will about how Americans see their place in the world ;)
[dead]
I worked straight through the dotcom, ‘08 financial crisis, and covid craziness. But recently haven’t worked in over two years, with nothing on the horizon. Definitely not normal over here.
Nearly 20 years then...
The GFC wasn't good? It wasn't good for the general economy, sure, but given the specific industry we are talking about that was one of the best times ever. Money was being printed hand over fist building software that did nothing more than emit a fart sound. The opportunities for us with the 'picks and shovels' were endless with everyone trying to strike gold at the App Store gold mine.
It took a bit for the VC world to heat up and a lot of big companies had laid people off.
The iPhone app store mania is not at all representative of the state of industry as a whole. Most software jobs did not look like this, at all.
It is unclear what you mean. It is true that most software jobs were not writing iPhone apps, if that is what you are trying to say, but those were good jobs. Those who didn't have good jobs for a moment were quickly scooped up by investors trying to create the next big app. It was a feeding frenzy out there for tech workers, despite much of the rest of the economy faltering (agriculture also did very well during the GFC, to be fair).
Maybe you are trying to say that there was some grey beard Atari programmer out there who refused to start writing iPhone apps and couldn't find their dream job banging bits on the old 2600? That is likely, but it is equally likely that they never found that job since either.
Correct. This is also the reason why job postings arent as high as they were during and after COVID. During COVID we had “the great resignation” which meant that people were leaving companies to find better pay after 2020 so companies had to “overhire” to account for the poor retention.
This is what a buddy who works at a major consulting firm told me about the hiring trend, we are returning to all the pre-COVID hiring norms.
This is the same with ANY depreciating asset or currency? Can I wait a year to buy it? Will I get 10x the return by waiting a year? Will my competition have a moat if I can move 10x faster next year? can I save money now?
I think this is a cope. It is not only the tech sectors facing poor labor market, it's everything. People can't find jobs
Idk why you're downvoted. Practically everyone i know is unemployed, underemployed, or failing to find a new job to replace one they don't like/isn't paying enough/has bad prospectus.
Lucky engineer types like me in the last group, but many of my college educated friends in that middle group. Shit is rough
I'll need to look at the underlying data, but it seems that the job "classification/category" should be more important than the raw number. A boost in lower paid/service industry jobs does not mean that there wasn't a loss in a category where AI can be more easily dropped into existing businesses.
Data : https://www.bls.gov/news.release/pdf/empsit.pdf
Average weekly and hourly earnings were up in May [1], though “real average hourly earnings for all employees decreased 0.5 percent from March to April, seasonally adjusted” [2].
Nominal wages being up rejects the hypothesis that folks are being downsized into lower-paying roles.
[1] https://www.bls.gov/news.release/empsit.t19.htm
[2] https://www.bls.gov/news.release/realer.nr0.htm
It's quite frustrating that they track average rather than median wages. As wealth inequality increases, average will be less and less representative of worker health.
They do track median and 25th/75th percentiles.
Do you have a link? There's no obvious place to look for this. All I can find is "Median usual weekly earnings of full-time wage and salary workers by sex, quarterly averages, seasonally adjusted". Which is not trivial to extrapolate into a real annual wage. And it's not clear how "quarterly averages" and "median usual weekly earnings" compose....
Old joke: Bill Gates walks into a bar. On average everyone there is a millionaire
I more like my high school math teacher explanation: "Put your one hand on dry ice, another in the boiling water. On average, you feel warm and cozy"
Except dry ice is -100 and boiling water is 200 so you feel hot and not cozy.
You're just adding the temperatures, you then need to divide by 2 to get the average - so it's actually 51.35 fahrenheit (10.75 celsius).
American education
Gold
That's a really old joke. :)
Here's a new one. Musk goes to a concert. On average everyone there is a billionaire.
The median would be interesting than the mean, as you can be hollowing out the middle, leaving more low-wage workers and a few very highly paid ones and the "average" still looks good.
Median is also up:
Median weekly earnings of the nation's 121.0 million full-time wage and salary workers were $1,235 in the first quarter of 2026 (not seasonally adjusted), the U.S. Bureau of Labor Statistics reported today. This was 3.4 percent higher than a year earlier, compared with a gain of 2.7 percent in the Consumer Price Index for All Urban Consumers (CPI-U) over the same period.
https://www.bls.gov/news.release/pdf/wkyeng.pdf
wonder how much of it is K-shaped. if is payouts for execs, capital gains and alike are boosting aggregate.
Median specifically avoids outliers at both ends of the continuum like that.
Always an excuse, huh? Both mean and median wages are up. Sorry this doesn’t suit your narrative.
What is the employment rate between the two quarters?
Edit: 4.2% (2025) vs 4.3% (2026).
https://fred.stlouisfed.org/data/UNRATE
In absolute terms AI is nibbling on a fairly small slice of the global pie of jobs - junior coders, lawyers, accountants, bankers.
The average person earns an average salary doing something very different to the stuff we on HN stress about.
key point, those 6 figures SWEs earn, they would spend to "local non SWE" economy.
your local datacenter does not care about local chickens or eggs, or private tutor, or pretty much anything at all. not even energy, it is has its own nucleaer reactor nearby. it is one-way economy from now on. you are only a consumer, not a producer. there is virtually nothing you (nor average joe) can provide that "datacenter" needs.
It doesn't have a nuclear reactor, it has two dozen natural gas turbines.
This is - quite literally - the fixed-pie fallacy, which has been thoroughly debunked by now. Read up on basic economics before commenting on this topic:
https://en.wikipedia.org/wiki/Lump_of_labour_fallacy
be careful with "real" hourly earnings. if they're using the CPI then it can be very misleading. One need only compare nominal wages vs nominal rents per square foot for the last 50 years to see that actual wages have dropped by 30% and yet the "real" hourly average wage calculated via the CPI is flat, falsly implying that actual wages have not dropped.
> if they're using the CPI then it can be very misleading. One need only compare nominal wages vs nominal rents per square foot for the last 50 years to see that actual wages have dropped by 30% and yet the "real" hourly average wage calculated via the CPI is flat, falsly implying that actual wages have not dropped.
Yes, if you get to pick prices of arbitrary items to compare against, it's easy to come to whatever conclusion you want[1]. That's why CPI uses a basket of goods, specifically to avoid cherry picking shenanigans.
[1] https://www.aei.org/wp-content/uploads/2022/07/cpi2022junea-...
CPI functions to obscure disproportionate price increases in essentials by lumping them in with cheapening commodities (e.g. electronics) that have a high demand elasticity. It's as much a choice to to use CPI to come to the conclusion things are fine, as it would be picking "arbitrary items" such as the essentials: food, shelter, medical care, and transportation that would paint a picture distinctly less fine.
>CPI functions to obscure disproportionate price increases in essentials by lumping them in with cheapening commodities (e.g. electronics) that have a high demand elasticity.
Right, because it's not "essentials price index" or whatever.
>as it would be picking "arbitrary items" such as the essentials: food, shelter, medical care, and transportation that would paint a picture distinctly less fine.
The BLS makes the constituent price indices available as well as their weights, so you can easily vibecode an "essentials" index.
shelter isn't just a cherry picked item. It's by an far the largest spending category, that and healthcare. I couldn't care less about all the other categories, personally.
>shelter isn't just a cherry picked item. It's by an far the largest spending category, that and healthcare.
The BLS agrees. That's why "shelter" is weighted 35.6% in the CPI basket, by far the biggest item.
> I couldn't care less about all the other categories, personally.
If we're going by what people "care" about (whatever that means), the basket would probably be like 70% gas prices, 20% grocery prices, and 10% for everything else. Empirically speaking, those two are far more salient politically than housing.
yeah, staggering. it's utterly amazing how much people care about gas prices. i can't wrap my mind around it. I mean, would people really rather be out on the street rather than get an EV or god forbid a hybrid?
Gas is misleading - petrol / oil would be better. I don't know if natural gas is going up, but if we mean pump prices of the liquid that goes into cars: that's also used (although with some differences) to heat people's homes - oil heat is extremely common in rural areas. This winter it cost me about 600 dollars in one month to heat a home in the Northeast I kept at 60 degrees except weekends!
Electricity is even worse.
Petroleum also impacts industrial goods prices, flight costs, etc. It's more than just a talking point for drivers.
Gas prices affect very nearly everyone in America. High housing prices benefit the most vocal middle class constituencies (ie boomer homeowners)
Rent is a bigger consideration for most people i know, but most people i know take public transit.
Also keep in mind: people's rents don't suddenly triple in the span of a month when the president starts another illegal war... The shock value of gas, uh, shocks makes it easier to talk about than the slow relentless squeezing of the working class via rent increases that outpace inflation and wages
> that and healthcare
Careful.
Much of the western world has healthcare that is free at the point of use and paid for by taxation; one cannot extrapolate consumer inflationary healthcare costs in quite the manner the USA does.
Remember, CPI exists to attempt to figure out how the value of a currency has changed over time. It's not a cost of living index, a housing cost index, a wellness index, or anything of the sort. It is not trying to measure how you are doing in life. It is not there to determine the value of labour, the value of housing, the value of food, or the value of anything other than currency. It is simply compiled to determine the difference in value between $1 yesterday and $1 today, so to speak. As you can see, you cannot use the currency itself to provide that measurement. $1 and $1 are visibly the same; except we know that $1 and $1 aren't the same over time. That is where CPI steps in.
If you want to understand something else, use the measures and data that are focused on that something else. Right tool for the job and all that.
Huh? Housing becoming significantly more valuable than in the past[1] does not imply that wages have dropped.
[1] Mostly thanks to computers/software making housing a more compelling proposition. Historically, people couldn't wait to get outside or over to the pub and would allocate their spending to support that, but now that they prefer to stay in to look at screens, they deem that place where they spend time worthy of greater spending instead.
IIUC, it's too early to taking "0.5%" as fact. Need to wait a few more months to see if any revisions occur.
What would cause an increase in the number of open lower paid and/or service industry jobs while simultaneously reducing the number of openings in tech?
World Cup / Summer.
typically unemployment rates are seasonally adjusted.
That annual US world cup season?
Exactly. Without any specificity of job categories, this chart and conclusion are unsubstantiated garbage.
AI can't flip burgers?
[dead]
It's crazy that so often I see articles, here on HN and elsewhere, where some pundit claims that there is no AI job crisis, AI isn't replacing any jobs, that layoffs are actually due to post-pandemic ZIRP overhiring, etc.
But then people who work in actual tech companies come in and explicitly say they are not hiring any juniors anymore specifically because AI is good enough to do most of what juniors do, and that senior engineers can now write 3x as much code, etc.
There seems to be a desire for a narrative that AI really just can't replace productive work, and that it's all a mirage. However it seems just like common sense that if an AI can do junior-engineer-level coding work, that a company has less reason to hire a junior engineer.
The decline in junior hiring began before ChatGPT had wide adoption, so AI is not a likely cause.
https://www.employamerica.org/labor-market-analysis/dont-bla...
The New York Fed has also released some research suggesting remote work has been a major factor differentially affecting early career workers.
https://libertystreeteconomics.newyorkfed.org/2026/06/remote...
"But then people who work in actual tech companies come in and explicitly say they are not hiring any juniors anymore specifically because AI is good enough to do most of what juniors do, and that senior engineers can now write 3x as much code, etc."
If you want an anecdote: the media company I work for just started hiring interns and juniors in software career tracks again after a lengthy hiatus.
I have a suspicion that Twitter laying off so much of the software staff was very influential for people with hiring abilities. The company didn't crash, and they (relatively slowly) began shipping new features again. I think that coincided with the pandemic-era overhiring, and we've been working against that combo ever since.
Every now and then, I actively try to make an LLM replace my tasks, and fully do greenfield projects I would accept- I don't see it. It's very good, no doubt. But I have or have been given the project parameters, and just like with a junior, failures in communication inevitably lead to breakdowns in execution.
It does your job, but not completely autonomously so you don't see it?
It requires a lot of guidance, luckily, I mean thank god otherwise we would be goners. The job itself hasn't become easier, but it did change. If you come across failures you update the spec, the guard rails, whatever you use to guide it. It's not a "it produced bad code and now it's forever useless" type of situation.
[flagged]
You just did. But it's OK, I understand.
I thought 80% of those layoffs were bougie HR jobs
I hate to be "that guy" but you're crazy if you don't see AI being able to do greenfield projects you'd accept.
I mean, in your defense, last year I think you would have been right, but right now? Codex rocks, as does Claude. I am literally making money shipping a greenfield project to a customer right now. I'm basically a cheap consultant that is incrementally adding features to make something exactly what they want for way cheaper than it would be to do it the old way.
There are hiccups, outages, things to fix etc. but the customer is happy with the output, and the reduced price means they get bespoke solutions rather than some BS one size fits all SaaS app. Then my job is maintenance and effectively "ITSM." Which kind of sucks in some ways, I miss writing real code for real projects, but this is the future going forward. If you want something for your business, you'll generate it rather than pay for it and for now at least, getting beyond localhost requires someone who knows a bit about computers or is willing to learn. Most small businesses aren't willing to learn.
Now, to your point. Is the code all that clean? Nope (and in your defense sometimes I read through the codebase and shudder)... but who cares? Like, for awhile I would go through and frantically edit it, but why? It worked. Not only that, but there's going to be a new model in 3 months or whatever that can clean it up and make it less shitty. I've literally done that a couple times since I started doing this in January.
The customer ain't reading the code. They don't care as long as the the functionality works - that's what counts. The gazillion tests I have keep it stable as I push code, and the CI/CD pipeline removes a ton of the ass pain I'd have without it.
The biggest thing I'm worried about when it comes to clean code and good design is trying to make sure I keep the token count down on these projects so I can actually do meaningful work without burning through a week's worth of tokens in a single day. That, and I like to try to keep a sort of architectural bird's eye view on what's happening...
Like, I'm not sure what niche of the industry you're in, but for the stuff I'm using it for, stuff is working really well with LLMs.
>Now, to your point. Is the code all that clean? Nope... but who cares?
Typically, this is not the type of phrase that is said right before everything goes extremely well
I assume they are working on low stakes software. Does it really matter if you use an LLM to code a scheduling app for a hair salon or veterinary clinic?
Meh, stakes don't matter.
So much of the financial world runs off excel.
Think about all the geneticists that complain about excel re-writing DNA sequences.
It doesn't matter if its high stakes or low stakes. People use the software that generates the results they like. Not the software that is "correct to use".
This is it.
So to try and understand your position - you are hired by a small? Company in which sector? And you built an app that does what? And I think most importantly - how did you find the gig? Were they explicitly hiring a AI capable person to do X?
I run my own thing since the start of the year. I’m building little tools for an industry I’m highly familiar that needs very specific scheduling software, data tools, tracking tools etc. My old job was a boring as a government bureaucrat.
I started this by doing some work for an old employer that asked me to start by modernizing an excel spreadsheet into an app I made for them like 10 years ago? They kept asking for more though since, and they’re my biggest customer right now. Which is good because I only have the bandwidth for like one more place right now.
I’ve had a few sort of one off things with other places? But I’m working on getting another company in the same industry right now and I’ll be able to adapt most of the code I’ve built here for another company if they end up deciding to use me.
But my biggest value to companies is “I already know this industry extremely well.”
I'm curious, did you choose "blind pilot" as your username before or after your adoption of LLMs for projects?
I am a blind pilot lol, but I’m not going to dox myself further
Since this comment began so rudely, I decided against reading the rest. All the best.
You do you, I literally say why you would have been right a year ago, but “so rudely” is a pretty funny stretch.
There's no rule that a phenomenon needs 1 cause - the decline could've been junior cuts to start after '22 since they have less seniority and impact, followed by a substitution effect after broad LLM adoption. I don't doubt that remote had an additive effect on productivity via mentoring etc as well.
> https://www.employamerica.org/labor-market-analysis/dont-bla...
I would have preferred someone who uses a null hypothesis and statistical tests.
> There seems to be a desire for a narrative that AI really just can't replace productive work, and that it's all a mirage.
Yes, and juniors aren't known for their productive work in the beginning. That's not their purpose. Their purpose is to do the mundane work, because it is important for them to become less junior and more senior.
That is robbed of them.
Which in 5-10 years means the need for senior developers is gonna shoot through the roof.
Exactly. Now most new hires are apprentices, before they might have been hired to do chores
people keep saying that assuming that in 5-10 years AI won’t be as good as senior engineers
If that is the case there is a worldwide crisis. It means that nearly all knowledge work will be gone.
What hubris on HN, to downvote this very reasonable comment.
A year ago I had no use for any LLM-driven tools, and today I can do my entire job without writing a single line of code.
If that happened in a year, what can we expect in 5? I have no idea what my job might even look like by then.
What makes you think that? AI isn't as good as any human in any field. Why would it be capable of replacing anyone at all in 10 years?
5 years ago AI was a joke, today it is a major industry tool.
>What makes you think that? AI isn't as good as any human in any field.
Not even chess-playing? More applicable to jobs though, they're arguably better than some juniors.
What are you talking about? AI is currently better at the things it can do than the average random person on the streets.
The need for seniors isn’t going to “shoot through the roof” if they are using AI.
If senior engineers are even 2x more productive with AI, then it’s like there are 2x as many senior engineers.
Most likely, seniors will be 10x more productive in 5 years using AI. This outpaces the retirement rate.
All the software engineers we need for the next 20-30 years are already in the current workforce.
Only way juniors can rise to the level of seniors will be through independent projects, long unpaid internships/apprenticeships, etc.
The industry will now have heavy gatekeeping built in.
> All the software engineers we need for the next 20-30 years are already in the current workforce.
There's no way of knowing what the industry will look like decades from now. Even assuming the prediction that seniors become 10x more productive, that would mean software becomes much cheaper to produce. Does that induce enough demand for additional software that keeps employment levels high? Could be, who knows.
Alternatively, maybe software hits a saturation point where there just isn't as much new ground to cover and employment levels crater. That could happen too.
Ok so even less demand for software engineers? All the software engineers that the industry will ever need again have already been created?
... and by the year 50, they will be a trillion times more productive?
>But then people who work in actual tech companies come in and explicitly say they are not hiring any juniors anymore specifically because AI is good enough to do most of what juniors do, and that senior engineers can now write 3x as much code, etc.
Yup, that's reflected in the data as well, no need to invoke "vibes" or whatever.
https://www.economist.com/content-assets/images/20260516_EPC...
https://www.economist.com/content-assets/images/20260516_EPC...
The likely explanation is that there's job losses happening in some sectors, but it's made up for in other sectors.
Can confirm. From my experience, since December 2025, state-of-the-art models can replace a lot of need for junior, medior or senior developers.
This meant that, in my example and given the same scope, one junior person with AI could easily handle multiple projects in 2026 that would require 2 or more medior people as recently as 2025. The only reasons for involving multiple people were time, learning opportunity, or increased scope - not skill.
If we paused AI progress today, and ignored the option of adding agents recursively, the current state of things still provides several alternatives to any honest & informed manager:
- Altruistic or optimistic: Keep all current jobs, and move the scope until there's a need and/or funds to hire more devs;
- Pragmatic or neutral: Keep the current jobs, but not hire anyone else as long as the scope is balanced against time (our current situation);
- Cold or pessimistic: Lay off anyone who doesn't fit the leanest model at current (or acceptably lower/higher) scope.
I don't know whether this translates into a "job's crisis", and I can only hope it won't. But hope is not a career strategy, scope is not inexhaustible, and the political, economic and social pressures are quite strong even if you removed AI.
They probably weren't hiring any junior coders to begin with.
> actual tech companies
Are you talking the big 10? Or "tech companies" in general?
> AI really just can't replace productive work
Okay. Show me the productivity gains. Those are measurable. Why is the "AI is ready" crowd never prepared to show this?
> that if an AI can do junior-engineer-level coding work
Then you have no competitive edge and most of your output probably cannot be copyrighted.
If you’re so confident, show us some data to debunk the article. You have a weird chip on shoulder, but no economic evidence to justify it.
I think that's coming at it backward. The article lacks data. Nobody should be expected to run around writing the article that's not there. As it stands, it's unsubstantiated junk.
What do you mean the article lacks data? Did you miss the BLS report?
There's no report. This "article" was three sentences and a chart that doesn't even break down crude occupational categories, let alone anything more specific.
Essentially useless.
The one with a weird chip on their shoulder is not the parent I can tell you that much.
> However it seems just like common sense that if an AI can do junior-engineer-level coding work, that a company has less reason to hire a junior engineer.
I mean, if we want to not talk about economics that’s fine, but can the AI actually do junior work at the same price? What if we don’t look only at quarterly reports, and instead include the value of having people knowing about the business having to explain it to others, who then learn it and can improve it over time?
I think it’s clear the AI is strong, there’s no doubt about that, but that’s not the whole picture.
>I mean, if we want to not talk about economics that’s fine, but can the AI actually do junior work at the same price?
Even if we assume it can then not hiring Juniors still doesn't make sense - where will seniors come from in the future?
That's a problem for the next CEO of course
This is a collective action problem. Its too easy to claim the answer is "we'll hire our competitors juniors when they become seniors" and then every company wants to do that and not train their own juniors. Soon no one is willing to train juniors because theyll just leave for the companies that dont train juniors.
The point is not whether it’s a mistake to not hire juniors, but whether it’s actually factoring into the hiring/layoff decisions at tech companies. Many claims are being made that no companies are actually changing hiring/layoff decisions on account of AI, and are using it as a distraction to avoid admitting their mistakes over hiring. That may be true, but many managers and execs actually do seem to consider AI a sufficient replacement for much of their engineering team, and are stalling hiring/prompting layoffs because of it.
It's not so much a question of whether AI is strong or not. It's a question of whether the tradeoffs (theft of intellectual property, coal burning, lack of transparency, stealing water, rising energy prices, global surveillance, etc.) are worth the outcomes. It's not even a serious question.
If AI was truly strong, we would be seeing signs in the job market. And we would certainly be seeing a lot more subscriptions and demand for these services. For most people, AI does not improve their lives. For a lot of them, especially younger people, it makes their lives much harder and sadder.
> If AI was truly strong, we would be seeing signs in the job market.
Recalling past business cycles, the job market can look "strong" with low unemployment, even while a lot of creative destruction is going on. Look at the 2004-2009 cycle, where the economy looked strong until some time in 2008 even though a lot of money was being wasted on bad projects.
It's only when the tide goes out that you see who's swimming without trunks. So unemployment doesn't typically rise until the recession part of the cycle.
> include the value of having people knowing about the business having to explain it to others, who then learn it and can improve it over time?
It's always been difficult to put a number on that value, which is the problem for the MBAs running the show. There's no number on the P&L assigned to tribal knowledge, and improvements that can be made by those with that knowledge and experience within the business.
It's a mistake that businesses keep repeating, over and over again, yet never actually learning the lesson. And now the industry is going to repeat it again, until there's enough pain that they realize that lost value and start rehiring again.
the job description of a junior engineer can change. junior engineers can use AI to make themselves more productive too
In my experience, AI raises the ceiling but also lowers the floor. It can make experienced people more productive: they can vet the output. The juniors? Not so much. I've worked with guys who write prompts that are literally "fix it." The result is about as good as you expect.
"Productive junior" is vastly different than a productive software developer with a couple decades of experience. What they produce will differ in quality significantly, AI or no AI.
well said. my wife's company has been having layoffs particularly from the overhiring during COVID. They went from 10 billion down to 1 billion thus far.
And now it becomes so good, that you want to have it. Which means companies have to deal with the budget increase which they will recupe.
How? By making a 5 person team a 4 person team + AI.
If i think about my co-workers (not excluding myself) from last 15 years, there was always someone you would accept just because it was better than not having that one person. If i can now replace them with more tokens/better models, man i wouldn't hesitate (of course i know what this means on a person to person level :/)
Not just juniors, honestly the hiring slump that started in 2022 keeps going on with no signs of reversion.
The AI job crisis is large tech employers firing workers to free up cash for AI capex, and those workers competing with everyone else in a competitive job market.
I’ve been unemployed and actively looking for a job for about 6 months, the longest stretch of active job searching I’ve ever done in my career. Several close friends who work in tech or tech-adjacent fields are in the same boat. Anecdata on Hacker News or LinkedIn tells the same story.
A chart showing “total number of jobs” is not meaningful. I took a temp job in a metal shop to make ends meet while I wait for the endless rounds of interviews I’ve now gone through with 4 companies. It pays less than half of what I was making before, and I am barely making ends meet. It’s not sustainable, even though the pay is more than fair for the work.
There are also a lot of job openings for home health care workers, or seasonal resort workers, that used to be filled by immigrants. Those jobs are not going to be taken by any of the engineers who just got laid off by Meta.
I have the strong impression that people who write articles like this are very disconnected from the reality of the economy right now, and that their curiosity ends at the line on the chart they cooked up to make a contrarian point.
I am very sympathetic to your situation, and fear something similar happening to me. But the article is addressed to investors, who are interested in aggregate labor demand declines.
"A chart showing “total number of jobs” is not meaningful."
It is meaningful in answering the question being asked, which is whether the hype around labor displacement, which has been growing for nearly five years now, is actually occurring in a way that would justify some of the higher valuations for AI firms.
The "AI jobs crisis" is generally understood to mean a sustained downturn in demand for all labor due to AI substituting for labor across a huge range of tasks.
Same boat, used to do backend infrastructure but have now been doing basic IT support at a non-tech industry; on paper I'm still employed, but between making not enough to make ends meet anymore and not even being fully engaged with the work, I feel effectively unemployed.
That's true! In 12 months, I got two contracts: a 3-month fixed-term contract and a previous 10-day contract role. Let's say that the landlord decides to sell the place I'm renting, I won't be able to rent another place due to unemployment.
I run a job search site and I don't see a crisis in terms of job openings even for SWE - but there is very clear signal that AI is deeply getting embedded in every SWE job.
https://corvi.careers/blog/global_software-engineering_jobs_...
Your own site shows there's very very little demand for entry level SWE which mirrors all the other data we've seen.
BTW, cool site!
Thanks. Yes, entry level roles appear to be under pressure, and middle-management roles in the Bay Area may be seeing similar pressure. (anecdotally) What is less clear is whether these are early warning signs of a broader shift, or simply cost cutting measures to account for AI spend.
IMO, AI coding tools have materially improved SWE productivity compared with two years ago. The open question is whether demand for new products and services will expand enough to absorb that productivity gain. If demand does not increase correspondingly, then it may only be a matter of time before we see meaningful job market pressure, starting with SWE.
In the news I hear about tech layoffs, but I am having a difficult time trying to hire high quality senior SWE's. Perhaps it's not the senior's that were hit by the layoffs, or I'm just looking in the wrong places.
The big job sites feel rife with unqualified applicants and outright fraud, leaving me frankly unsure how to proceed.
It feels like a crisis of candidates-to-job-openings matching, and AI has made it much worse.
Perhaps its because a good chunk of us looked at wtf is going on in tech right now and said "nah, I think I'll go be a farmer now."
I see dozens of senior postings, i apply for them, i get automated rejections. Wash rinse repeat for 6 months.
Is your job remote friendly and what's the pay?
https://jobs.ashbyhq.com/rivora/c146a25b-dddb-4ac0-a8c1-c8b3...
Just a minor suggestion if applicable: My site didn't pick up your listing - Corvi has a direct integration with Ashby. I recommend enabling all of job boards in integrations to get your posting to job search engines beyond popular ones.
Oh, I'm on the job market, unemployed and probably wont apply becasue of the contract. That's probably your reason for not getting the quality of candidate you're hoping for. that hourly range goes pretty low and doesn't give confidence that a potential employee won't get low balled
Thanks for the tip! Ironically it’s the opposite in this case; where we’re willing to exceed the upper bound for the right person.
Your point about the good people self-selecting out makes sense.
Have you tried to verify if those job positions are real? Try to apply yourself and go through the interview process to understand what's going on.
I'd correct that to forcefully embedded, like unpaid on-call.
I am a manager with former FAANG experience followed by mid-size comanies. Most people here saying there is no AI job crisis dont know what they are talking about and talking out of their ass.
I know plenty of people looking for job for months at this point. I am looking for 3+ months and have a few leads finally but nowhere close to offer. Remember, people have to live in these high cost GEOs (SF, Seattle, NYC and others). Since COVID, the US economy is effy at best with housing prices doubled, rents and household expenses doubled. Just for something simple it costs $200 subscription like mowing the lawn. Pulling stats out of some dashboard that a rando made is not how people live day to day life. Remember, people lie with statistics all the time.
So yes, there is a significant job crisis. Because Ycombinator invested in AI companies this narrative is suppressed here and moderator DANG (one of the worst) heavily moderates anything critical of AI and YC.
@BLKNSLVR - What does account creation time have anything to do with my comment? You are demonstrating classical case of logical fallacy.
I have other account but I don't want to post with that or have it banned by censorship king DANG.
If you don't want to believe that is fine but don't make surreptitious accusations.
This account was created an hour ago.
Personally, this makes me suspicious of calling out mods for censorship. This, in turn, makes me suspicious about the truthiness of any of the information presented.
I'm not agreeing or disagreeing with any of the opinions or experiences expressed.
> So yes, there is a significant job crisis.
And the link echoes that there is a job crisis. But not an AI job crisis; rather a COVID frenzy fallout job crisis.
Anecdotally, there's an AI job crisis for juniors right now
Yea, we're basically not hiring anyone that isn't a senior developer already. That's going to be a huge problem eventually but not my problem to deal with.
My best advice for folks that want to get into software now is be willing to do it cheap for awhile and then jump once you've developed some skills. If you were getting into this industry for the money you're properly fucked and I hope you didn't load up on debt. If you're passionate about building stuff there's still room but the path forward is a lot murkier.
This has less to do with LLMs than people think.
The reality is most firms are running out of projects to take that make economic sense.
Note: ECONOMIC SENSE. This has nothing to do with refactoring for the sake of refactoring. Its all to do with earnings growth with respect to the cost of capital.
Me and most of my friends from school started in low paying consultant jobs, in fact most from my class dont do software at all only 5 of us had jobs after education ended, this is like 12 years ago, it was also hard back then to get a job, maybe people forgot or something.
This.
I have been out of the tech startup labor market for many years. But when I was doing hiring in a startup around 2017, we had to put specific, conscious effort into accepting inexperienced candidates. The default mode for hiring is finding like-minded individuals whose experience overlaps with the existing team.
Using AI as a scapegoat for hiring freezes or layoffs is just more FUD. Like all propaganda campaigns, the purpose of fatigue is to induce an amnesia effect.
This has been my experience as well. I tried my hand at switching jobs in 2018 and it was a tough market. It feels like people don't have any concept of transferable skills, and just want a clone of someone who just left and can pick right up from where the other person left off, with no on-boarding period. Nothing seems to have really changed in the intervening 8 years.
> is be willing to do it cheap for awhile
Then you might as well work for yourself.
> getting into this industry for the money
I can make more money doing HVAC but I'm tired of being on hot roofs.
> the path forward is a lot murkier.
If you're just here for the money go somewhere else. If you're here because you love computer science then ignore these people and do the work. If you can't find a company get a dayjob and do it for yourself.
> If you're passionate about building stuff there's still room but the path forward is a lot murkier.
Definitely feel the murkiness. I've been programming as a hobby for over ten years and only recently started wanting to do it professionally. I'm actually wondering if there's a path for me.
What people don’t get that traditional software jobs are gone.
There is no need for developers, testers, PMs, TPMs, devops, leads anymore. Communication burden and structure imposed by these roles is too high when you have AI. It doesn’t make sense to tell somebody to do something and for them to type it into the AI. You can type it into the AI yourself without wasting the time.
There is new job - software creator. Think of it as a single person replacing a team. This job requires different mix of skills, and different level of autonomy. Hiring needs to be adapted to this really, and people need to adapt. Some will shine in this new world, and those who still narrowly think of themselves as specialists in the specific old role are going to be jobless.
same applies for seniors as well. ther isn't much distinction of senior vs junior human dev (as in cost and efficiency) compared to AI-dev (cost and efficiency). more so, at current imrpovement rate. in couple more years you would not need seniors anymore either.
As long as architectural decisions have long-term costs and you need human taste and judgement to speculate on what business needs will come about in the next few years you'll still need human engineers.
We replaced most of our juniors and seniors at my company with AI.
We only have like 2 seniors down from 30 and they are more productive than ever.
I expect with Mythos we can get this down to at least 1 highly paid senior engineer and the rest a swarm of agents they are managing.
What exactly was your company doing such that the concurrent development of features and infrastructure by 30 people is now done by 2? Each engineer is monitoring 15 Claude swarms?
We just have regular web, backend, search, ingestion, frontend and mobile codebases.
Turns out we didn't need these engineers and you can just have one or two strong seniors on this.
They are more productive than ever.
[dead]
This seems pretty unlikely. If it turns out to be true then you don't need a junior or senior dev you can just get a random person from the street and they could do the job.
that is what CEO of NVIDIA is telling everyone. "everyone is a programmer now".
I don't follow. Aren't you arguing something more like "no one is a programmer now"?
yes. software engineer as a paid profession is going away within couple years.
when everybody can just press a button in elevator, you do not need dedicated paid person to do it.
Finally, everyone can code. Thanks to all the programmers for their efforts
This is a good point. So many stories on Reddit of college grads in computer science unable to find work despite being qualified.
tbh, it was always kind of difficult to get a job as a Junior engineer, I had to work for free for almost a year to get the job then slowly grind my way up for higher salary
It seems to me that the hiring crisis started in 2022 and never ended.
I haven't been on the market for long luckily, but as an independent consultant I went from getting 2 to 3 contacts per week to none per months.
It's really funny that the first thing in the article is a graph showing a long downward trend with a tiny, brief uptick circled at the end and then the entire article appears to be written about that.
I don't see an uptick. Demand has been steady for the past two years after falling from the post-pandemic frenzy. That fits with my personal experience of the tech industry over the past several years.
People are right to point out that hiring is nothing like the post-pandemic years, but it's not clear that it's any tougher out there than, say, 2018. This is from the perspective of someone with a lot of industry experience, though. I can't speak directly to the experience of junior engineers.
True, but coding agents have only been practically useful since early 2025. Jobs are flat or a bit up since then.
Bold claim to say "no signs" based on non-contextual numbers. I think I recall somewhere that most jobs added were in healthcare.
>> The May jobs report reinforced this with nonfarm payrolls jumping by 172,000, confirming that there are no signs of workers being replaced by ChatGPT.
The shift toward healthcare employment is a very long running trend driven by the greying of the Baby Boom generation.
https://fred.stlouisfed.org/series/CES6562000101
It's not just that. It's anyone who wants to live longer and look younger. Pretty much anyone with the income to afford it. As society becomes wealthier , it means more $ spend on elective procedures and healthcare overall. Wellness clinics are a huge deal now.
Yeah I'm starting to think the BLS needs to do "Boomer aging adjustments" to jobs numbers, in a similar manner to how they do "seasonal farming adjustments". Until we get through the Boomer population bulge, healthcare is going to keep adding jobs for quite a while regardless of how the rest of the economy does, but that doesn't necessarily mean the overall labor market is healthy.
If you leave out healthcare, 2025 had massive job losses overall, with Boomer bedpan cleaning bringing the net number up to just above zero.
If there is an increased demand in healthcare jobs that will increase wages in healthcare which will pull people out of other jobs and into healthcare in a healthy labor market. I’m not saying whether or not the labor market is healthy, but this adjustment wouldn’t help you figure that out.
Also Gen X isn’t that much smaller than the boomers, and millennials are the largest generation ever. Plus all generations aster the baby boomers have fewer children per couple to take care of them, so demand for healthcare jobs isn’t going to drop anytime soon.
> the BLS needs to do "Boomer aging adjustments" to jobs numbers, in a similar manner to how they do "seasonal farming adjustments"
You’re comparing a low-frequency trend with a high-frequency cycle. The latter has lots of data to characterize it. The former may be secular or may be a slow cycle; nobody should be adjusting for it in the base data.
It’s so frustrating that someone with a title of “Chief Economist” puts things like this out there.
Job numbers get revised every month, in a negative direction.
New grad unemployment is high and trending higher.
New jobs exclusively are held up by addition in healthcare industry, almost every other sector is seeing some negative movement.
A lot of job openings, a good chunk of them, are just fake jobs where the company has no intention of filling them.
Pretty bold for someone to ignore all of that and come up with a claim like that.
There are upward and downward revisions this year at least, it seems like a mixed bag with many policies juicing the economy. [1]
Job opportunities differ by state and de-growth hostility to business policies and crony investments. Where I am, layoffs and offshoring continues. I hear new grads are increasingly opting for the skilled trades, which is interesting given they aren't getting use out of their degrees.
[1] https://www.bls.gov/web/empsit/cesnaicsrev.htm#2026
> There are upward and downward revisions this year at least, it seems like a mixed bag with many policies juicing the economy
Since Jan 2025, the only two upward revisions have been in the last 3 months. So we’re both correct here. This year (3 months of data), has been a mixed bag. But on a longer time horizon (last 6, 9, 12 months) the revisions have mostly been negative.
Yes, I hope I didn't contradict you, the economy has been shite. Was just trying to leave a positive note on recent developments that will continue, God willing.
Well... if you think in terms of a society spending its people on doing various things, spending more people on healthcare could be a good thing. It means we're getting food grown and stuff made with fewer people, so we can spend more people on making sure that everybody lives, and lives healthier.
Healthcare is a huge sink if you have the income to afford it . People will spend hundreds of thousands of dollars on various protocols to look younger, beat aging, and so on.
Workers taking on new/different roles isn't the same as being replaced. Workers have been taking on new/different roles since at least the advent of agriculture, so that's nothing new. Being replaced would be something new, but the data doesn't support it.
The most glaring gap in this "analysis" is that it's too early to tell. I don't understand how people can be certain in both extreme viewpoints.
How much time has to pass before we can dismiss the widely cited predictions by the media (and also many popular bloggers and podcasters) that AI will lead to mass joblessness? Those people need to at the very least acknowledge that their forecasts are so far wrong.
To answer your question: It depends on your claim. All claims of mass unemployment thus far are evidently not true. All claims of mass unemployment in the near future: We'll have to wait and see.
The prediction that AI will lead to mass unemployment is largely a widespread fear more than an observation.
AI washing: On the one hand, you have periodic mass layoffs that could be related, but also could be "AI washing", with the companies citing AI to sound progressive when, in fact, they just over-hired.
Multiple independent sources find no broad signal. (Anthropic's labor research, Yale Budget Lab, Apollo's chief economist Torsten Sløk)
I'd still like to know more about the junior situation, though.
Both sides should take a deep breath and stay away from predictions.
Both sides of what?
Hum... That's a huge conclusion from a single graph, I expected to see more data.
The jobs openings / unemployed close but a bit larger than 1 is the expected value when nothing is turning the economy completely upside down right now. It means there is no new shock that is too fast for people to adapt.
What is surprising, yes, because I could easily name half a dozen shocks that I would expect to see there. But it seems US people are adapting quickly right now, or things are canceling out.
First of all the May jobs report was mostly in temporary workers possibly due to the World Cup. Second as already noted the jobs are mostly in healthcare. Third job openings does not equal employment. These numbers have been diverging for a while likely due to people holding multiple jobs. Also I believe the evidence suggests the job crisis is due to WFH and not AI https://news.ycombinator.com/item?id=48326721
A "job opening" is not a job. It's an aspirational advertisement.
Further, the graph shown is pretty noisy and I'm not sure the upward move which counters the downward trent is statistically significant.
BLS doesn't look at job ads when compiling "job opening" data. Their method isn't perfect (nothing in life is), but far more comprehensive than you give it credit for.
So what do they look at?
The results of them actually talking to businesses and asking questions that are more than "did you have a job ad posted?" You are hardly the first person to imagine that job ads aren't representative of actual job opportunities. Obviously they are going to put in effort to avoid those weak signals.
How many businesses were surveyed?
As many as was required to find statistical significance. This S in BLS stands for statistics, after all.
What are the flaws in this methodology?
Can a crisis exist within noise?
You can also look at the BLS unemployment rate. Its also low. The predicted mass joblessness due to AI shows no sign of happening
Advertisement or trying to fool investors that they are growing
The jobs AI crisis is me being stressed and overworked as management demands more AI, showy (but shallow) "look I made a skill" presentations grab attention, and my PR review queue grows every day as people generate more code than ever.
But to keep this out of a low-value vent, my experience has been that the _threat_ of it is there, but in my small corner of the world/industry, lots of layoffs that would have happened anywhere may just be categorized as "AI" layoffs, but the wild manpower reducing benefits aren't really there. The larger an org gets, the more of your job is dedicated to human stuff, and you can just get some of the code part done a little more quickly.
Would be interesting if we could measure how much effort is put into agentic coding harnesses, frameworks, and theory, vs labor saved using them.
Have never worked in my life like last year, and there's no signs of slowing down.
At least I don't have to review PRs, we have built a team around the concept of trust and not needing to check each other's work.
There's simply too much work to start nit picking in PRs, everybody's very talented and a good engineer you can trust. They ask for feedback when they want it.
And no, we're not producing slop, if anything AI has helped improve the codebase a lot and the harnessing has definitely raised the bar at all levels.
But sure as hell, the job sucks now, and I hate it. And it's not just due to the amount of work.
75% of the value on PR review is team visibility
We have built our team around not needing team visibility. We're 5 developers working on 7 products used by 5 different companies.
I *don't* want to have to know what others are doing nor how. This may work on teams that work on a narrower scope and lower rhythms, cannot work for us.
I'm a consultant, and had my first conversation about an AI clean-up job this week. I'm also just starting another gig analysing LLM output, my sell is that the analysis is hand-coded, as they weren't able to do it themselves with LLM support.
On the other hand, I'm just finishing an agent-heavy piece. After getting it set up, it's been some of the most mindless and soul-destroying work I've had the displeasure of in a while. This stuff will be near minimum wage in a few years, totally unskilled babysitting.
AI really hasn't been all that bad for work, by volume at least. I know where I want to focus my efforts though.
As a fellow freelancer where do you see that you could survive this into the future?
I operate in a bit of a niche, in terms of my clients. They're all typically people I've worked for full time in the past, or closely related. There are some fundamentals of that business I worry about, but I'm able to do things they can't, and I'm looking alright for the year. That's as good as I can generally wish for.
Longer term, I don't know. I'd happily take something more secure if it came along, as long as it's not childminding for an agent. Super busy and bored out my mind the last few weeks, the worst sort of work.
The expectation of linear presentation of change in a bistable system is gobsmacking here.
If this kind of argument were generally valid, it would imply that:
- all change neither accelerates nor decelerates, which is absurd, on the face of it;
- the initial stages of a deep change are always surface-visible; for instance, cancers announce themselves when they begin to gestate, rather than when they metastasize
- A few recent points of data of questionable significance outweighs a hypothesis with considerable support from reason, intuition, and other (unpresented) data. For example, the plight of recent CS grads, which _is_ new, and _is_ on graphs, just not the one the author here chose.
So, since these implied claims are self-evidently _false_, it means that the author would, at a minimum, need to provide an explanation as to _why in this one instance, these considerations do not matter_; for example, the author could have argued that the graph positioned at the center of their argument is the one to look at (as opposed to, say, recent CS grads,) but that _itself requires further argumentation._
It also does not account for the other obvious possibilities; e.g.,that there is a delay between the (as it were) lightning and its thunder; or that even strongly nonlinear effects would have shown up by now in the metric chosen; etc. But since these contributions were not included in the original post, I have no choice but to discount it.
Not really a question of linear versus non-linear change when we see, in aggregates, almost no change at all.
in, what, six months and like two points of data? No thank you, I prefer my thinking science-y.
I guess there are those who might say, "Well, hang on, give the crisis a minute to unspool..."
It's always 2-3 years away. 2027 was supposed to be the inflection point. That was 1-2 years ago.
same. I was expecting it to recover in 2025. but it only gets worse.
Well, I'm on a small two person team.. for the last few years I would have loved to have 2-3 people under me.. I've been underwater constantly, but we couldn't afford it.
Now I have no need for anyone from a coding perspective.. I can keep up with multiple clients requests with a breeze. I don't have to manage anyone. I type of my phone while I'm on a walk and work gets done for me.
So yeah... it's not good.
To me, it doesn’t make a lot of sense to hire fewer juniors because of AI. This assumes that the only thing AI can’t replicate is experience. This in itself already seems questionable, but most importantly what about creativity, relating to humans, common sense, etc.?
If you had to hire a concierge for a hotel, would you rather take a guest-oriented, quick-witted junior or an AI? If you could either take a Waymo or an 80-year old driver, what would you take?
I don’t believe hiring managers think that one-dimensionally. Roles are unique and in some roles experience is more important than in others. Plus, juniors already balance their lack of experience with lower salary expectations.
The much easier explanation for the anecdotes: companies are more cautious at the moment and if you only have a few positions to hand out, you rather take the proven hand than risking it on someone who hasn’t shown yet that they can do it
Btw. we haven't yet builed the whole Agentic AI layer which is coming and companies are working on.
A lot of people and companies working left and right building things which are obvious that they are coming.
We do not know yet how that impacts even more people.
And AI already removed certain jobs.
Exactly.
Much of the "AI job crisis" rhetoric was PR comms to manage conversations around corporate restructuring (even ZIRP is a lazy PR comms excuse).
Most decisionmakers by 2025 already agreed they didn't expect AI to have a significant impact on hiring [0].
I've pointed out the reasons ad nauseum on here but no one listens [1].
[0] - https://www.bloomberg.com/news/articles/2025-12-10/wall-stre...
[1] - https://news.ycombinator.com/item?id=47174561
For US SWE labor, off-shoring was and still is a big contributor. I do think AI is a factor too though. You almost get laughed at asking for headcount now. For junior positions in particular, there are close to zero openings.
AI has grown dramatically in capability since last year as well so I'm not sure 2025 data holds today.
It was difficult getting junior headcount in 2023 as well.
That said, I agree with you that AI capabilities have grown massively, but the only way you aren't burning tokens with marginal RoI is having someone with credible domain experience using Claude Code or Codex hand-in-hand.
Basically, no one wants to spend $200K TC on training a junior who will most likely be using an AI tool already. The math works out more positively at a significantly lower TC.
Someone here said that they get paid 50k to fix the ai code.
Maybe all the job cuts from ai were filled by fixers of ai output
Or maybe, no one ever heard of jevons paradox. Or maybe everyone ignored it and preached job apocalypse as risky but a high reward marketing tactic
AI Has Broken Hiring https://hbr.org/2026/06/ai-has-broken-hiring-heres-how-to-fi...
These analyses always crack me up. "There's 1 at least 1 job per person, why are people complaining" is the vibe I always get.
As though the decades of work I've put into my career means I should want a job as the hamburger man or working as a ditch digger. No shade to burger flippers or ditch diggers, but these are not jobs that I'm trained for, nor are they jobs that I remotely want to do. So for me, aviation expert, programmer, ML engineer, weird IT generalist, guy with a math degree who speaks a couple languages, and isn't exactly fully capable anymore, the idea that jobs in some other field (like an RN or an Oncologist or Electrician) are just something I can pivot into is just hilarious. It's such a shallow and ignorant take. Even the premise that all jobs are equally distributed and there aren't jobs that are more or less in demand at any time is a real funny take too.
I'm so glad I started working for myself, because honestly, seeing this dogshit analyses from supposed experts means I'll be able to keep making money for a long time just be actually trying when I need to think about something.
That little arrow on the right of that graph is wishful thinking.
I'm pretty sure I have enough tech debt built up to last through the singularity. And AI somehow is only increasing the amount of things I want to do.
Gedankenexperiment: If it turns out that the uptick in hiring is due to having to clean up after the first generation of vibecoded apps, is that really a net gain?
If America becomes wealthier due to AI, it will mean more follow-on jobs such as people using their newfound fortunes to remodel their homes, consume more , vacation, etc. Companies will expand and hire. All of this creates jobs even if AI may also destroy some jobs. The net result is more jobs. There is a huge market for upper-middle-class people in their 30-50s to look younger. This means more health clinic jobs and demand for pharma.
Theoretically, that's true. In practice the gains were limited to the wealthy who use that money to fund increasingly deranged start-ups* until the market crashes. Then rinse and repeat. The failed start-ups effectively waste any increased productivity leaving everyone about where they were before.
*They fund these start-ups to get a good return on investment so they can get even more money. As the economy overheats the number of places to invest with a reasonable return falls so they are left with the high-risk stuff to invest in. I'm not sure what they want the money for, though, since they could already afford most of the things I would find useful...
I think funding deranged startups is a type of consumption, and it does get money back into the economy the same way funding a remodel does. Maybe the reasoning to do so is different, and maybe the deranged startups add more capital than a bathroom remodel, but then again, bathroom remodels probably tank less often than startups, and worst case you can always read a magazine in your new bathroom. hard to do that in a dead startup
https://fred.stlouisfed.org/graph/?g=1WPjJ
Wow, zooming out really puts the 2021-2022 hiring frenzy into perspective.
Sort of, that it crosses so many lines makes it seem like it must be 6X, but it peaks at 230 based on a baseline of 100, so just 2.3X their baseline. Still a ton, but not as much as I thought at first glance.
The "edit graph" button reveals some pretty sweet ways to mess with the chart, yet I don't see an option to fix the Y axis at zero. Weird.
Maybe words work better?
There were roughly 28 months' worth of tech job postings within the 15 month period from July 2021 to October 2022.
If you change the baseline to a rough average of the last 2 years, 0.66, that ratio becomes 42 months' worth of job postings within 15 months.
I'd be curious to see what the numbers look like as a percentage of the existing tech workforce. Like, if there are 100 workers and the number of job postings doubled from 5 to 10, that's a huge deal. But if there are 1000000 workers and the number of job postings doubled from 1 to 2, well that's not a huge deal.
I don't see a software jobs crisis but I see a software industry crisis. The AI slopmageddon is upon us.
Pretty shallow analysis. Doesn't account or what sectors the new jobs are occurring or account for the extended length of unemployment many are experiencing or underemployment. If only he had asked chat GPT, it would have provided more context and nuance than a single chart.
If there's no jobs crisis, where are the productivity gains to offset all the massive datacenter spending?
If we need to spend just as much on salaries, while shelling out $700B/yr on AI, how does all that spending get paid for?
Wrote about this https://julienreszka.com/blog/robots-create-more-jobs-than-t...
I wonder when we’ll reach diminishing returns on new AI models. Haven’t tried Mythos or Fable yet but it seems like Anthropic is already priming itself for this by calling for a “slowdown” of AI development.
I think it's less of a technology problem and more of a money and investment problem. They called for the "slowdown" right when they filed for IPO. This allows them to both triage the money bleed while under the microscope as well as make their competitors look unethical and dangerous if they don't heed the warning to slow down before we lose control of A.I.
At least in cybersecurity, we already have seen diminishing returns with newer models.
At this point the harness/applayer matters more, as different models perform better or worse on exploit classes depending on the prompt, tuning, and various other parameters.
Of course, by the time HN hyperfixates on a topic, it's already been executed on and HN is too late.
Yea, in many usecases the tooling space is increasingly sophisticated context management such as fine tuning domain specific mappings into the model so that it is able to work directly with a compressed form of some data without needing to decompress into the context.
In larger models, these fine tuning techniques work more reliably/robustly. Because of this many usecases tend to prefer larger models. It is possible to work the same behaviour into the smaller model, but it requires more effort. But it's one-time. And smaller models are usually much cheaper. People make a tradeoff along this curve.
This is observed at few-B scale upto hundred-B scale. No way for us non-anthropic/openai to fine tune beyond that of course.
You can't draw any conclusions based on "job openings" without dealing with, or at least addressing, "ghost job listings". There are several issues here:
1. AI ATS systems have made posting jobs "cheap", such that too companies post jobs that don't exist (ie "ghost jobs") to keep up appearances they're hiring or just to keep people in the pipeline in case they hire. This is a huge waste of everybody's time and should be illegal;
2. The hiring process itself gets increasingly Kafkaesque. AI screening, automated online tests, unpaid take-home work, etc. You have to get pretty far until a human gets involved. 10+ years ago this didn't happen because people needed to be involved much sooner and that's expensive;
3. In a lot of companies, getting employees to interview people is unpaid extra work effectively. They say it's important. You might even get dinged for not doing it. But anyone who has done it realizes pretty quickly a bunch of people who shouldn't get interviewed are getting interviewed and management doesn't care, even though employee time is expensive, because you essentially have to "make up the time" so it's still "free";
4. Even if you go through all that and get hired, you get laid off within a year such that income isn't dependable and you end up wasting a ton of time on the job-seeking process itself.
I've been thinking about this recently and high-information is part of the problem. In years long gone, it was hard to reach applicants so you'd have a small pool of higher-relevance candidates applying for a job. Say 10 people applying for 10 jobs. The odds were better. It was less work on everybody's side.
But now you have 200 people applying for 200 positions. This wastes everybody's time but the problem is that companies have offset this by pushing filtering onto these automated systems. People still need to enter all their bio information, etc. So it's just much more inefficient inherently even if the job opening is legitimate.
You can look at the official unemployment rate, too. It's still low despite rapid advances in LLMs
The unemployment rate is cooked. It doesn't capture underemployment, people who want full-time employment and don't have it and people who don't make a living wage.
Long gone are the days when the vast majority of people just have one job. Now it's 2-3 part-time jobs because companies are exploiting a legal loophole where they only have to pay benefits for full-time employees. And then you have people doing "gig" work, which is often sub-minimum wage when you factor in expenses (eg driving for Uber and not factoring in car wear).
On top of all this we, across the Western world there's an increasing youth unemployment crisis. In 2008, entry-level jobs basically disappeared overnight and never came back. Well, that just got worse post-pandemic.
We're on the cusp of it. Models JUST got good enough to be better than the average white collar worker at nearly everything.
This literally is a brand new phenomenon; it's only a matter of time.
Literally the same thing people have been saying for two years. How gullible can you be.
Has there been anytime, when senior experts in a field, are consistently maintaining this time it's different? You can ignore this at your own peril.
No, you're wrong and have no data to back it up. I, on the other hand, bring valuable personal experience that says AIs are useless. I asked it to do X and it did X, but it did it differently than I would do it. This means it is fundamentally flawed and complete utter trash. If it can't do X exactly how I like it without any guidance whatsoever it is not ready for prime time and never will be because it lacks "intelligence".
If every company only hires seniors from now on, and if we assume that the number of senior level jobs is steady, why do I not perceive an increase in leverage like during covid?
The tech co I work for just massively scaled token usage after the copilot price changes. If there was a chance for ai to replace anyone, it's gone now.
We're still going through the post-ZIRP job crisis.
20 years ago, everyone was whining about losing their jobs to outsourcing. Seems like the more things change, the more they stay the same.
In the US there is a widening income equality gap[0], of which outsourcing is one plausible component in a large, complex system.
https://www.pewresearch.org/social-trends/2020/01/09/trends-...
My anecdote is there was a single wave of cuts and a bit of a shuffle, but for most part it's business as normal now
Maybe AI will finally be the tool that allows us to get rid of some of the people we have who do nothing more than push paper around. Maybe. But somehow I doubt it, at least not in a typical big corporate environment. And I have zero concern about us letting actual software devs go. Things will have to change pretty dramatically before we get that far.
AI will make those folks much more productive. They'll push a lot more paper around.
You mean the middle management? I have been in environments where they were almost literally made up of pencil pushers. Wouldn't be too sad to see them go. Only half joking, but it is written in jest.
Nope, just people who basically swivel chair information from one place to another. Useful in some way, should have been automated a long time ago, and yet persist.
But yes, middle management would qualify ;-). My manager seems spooked by LLMs. Loves to use them to write his emails, but seems to internalize that since they're doing his job for him at this point, his boss may figure it out.
What kind of jobs are these? Volume is just one factor.
I used to think comparisons of AI with the web where ridiculous, but increasingly it looks like they're not that dissimilar as far as how they change how we work. But as someone who graduated college during the bust there was a loooong gap between peak hype and things like online banking and e-commerce becoming standard.
Even if AI is an absolutely bubble, and SpaceX, Anthropic and OpenAI all cease to exist in a year... there's simply no way that AI has not fundamentally changed work. Even if I was forever pinned to the local models I'm running and the agent harnesses they use, I would never write code for work the same way.
But I lived through the rise of the web. I remember serving dynamic websites through cgi (which meant a new instance of an interpreter was spawned per user session). I vividly remember great JavaScript books saying things like "never use JavaScript for core functionality". I recall Java engineers saying Ruby on Rails was a toy and would never take off, that Python offered nothing over Perl and that "rich web applications" where never going to replace app native interfaces. I remember when the MVC pattern from the early Smalltalk days being dusted off and repurposed for web applications, completely changing how we designed software for the web.
And all of that is just software. It wasn't until the pandemic that ebooks replaced print books in share of academic library circulation (reversing a decades long trend of reduced circulation).
In my daily use of agents for coding and other forms of problem solving, while it is a wild accelerant, it's also clear we have not even started scratching the surface of how to think about building things with these tools.
I suspect we'll adapt to AI faster, but having lived through one major tech revolution, transforming work still takes some time. I'm not surprised we don't see an immediate jobs crisis.
Not to mention the completely separate topic that huge classes of employees were not and are not all that productive, so boosts in productivity don't imply lost jobs. That would require a boost in productivity combined with pressure to create concrete value with less, looking at the SpaceX IPO we're still a ways away from working about how efficiently we create concrete value.
This is a spin piece from a private equity firm. Hardly the most unbiased and credible source for this kind of reporting.
AI creates more work
Like others, let me call bullshit on that graph.
It just gives a general "job" going up, but
* how many of those are real jobs? I've seen the amount of scam/fake jobs go up a lot since 2022
* How many are jobs only for PhD with 20 years expereince?
* How many jobs for actual juniors / newcomers?
Just taking the aggregate is a bit scummy
Maybe give it some time?
Companies learned one weird trick, that's why you don't see job opening collapse. Most of the job openings listed are fake. Companies aren't hiring despite showing signals that they are; you can ask any engineer that's been laid off over the past year and they'll tell you the same story of how bad the job market is right now.
i swear people will always say it's just another fad until they lose their job themselves, might be a coping mechanism
the AI job crisis will be because the AI bubble will pop and th economic fallout is going to be tremendous. dotcom and GFC might not be as bad as what the AI bubble popping might be.
I mean, hasn't practically every single Trump-era jobs report been massively adjusted downwards after the fact?
[dead]
[dead]
[dead]
[dead]
Now that Claude Fable is out I’m sure it will finally come true /s
[flagged]
maybe look into industries AI is best at automating? like constant layoffs in Software? 160,000 people lost jobs in just 2026?
the claim "hey there is no AI job crisis", when previous SWE of 6-figures now takes job dishwashing in McDonalds + one more gig as Uber driver + food delivery gig is "job creation! now they have 3 jobs!". does not make any sense.
Why would you assume that’s from AI
You talk to any SE and it’s obvious we’re not running out of work to do since these tools became available
1. C-level says so 2. on-the-ground people indeed much more productive
but of course, it is not just AI. Software is consolidating and automating even without AI, that's the whole point of software.
Even if they're more productive, the amount of work to be done has also increased significantly. Not only is there so much work to be done on AI integrations but all of the other products have bug trackers exploding with new privacy and security vulnerabilities that need to be addressed.
Some of the security people I work with claim that without the productivity gains from AI, there is no way they could keep up with the bugs.
C-level does not seem uniformly convinced, and they have a lot of reasons to tell that story, even if it isn't really true.
don't know whom you talk to. I see people laid of left-and-righ, in FAANG, banks, startup, pretty much everywhere.
I hate to repeat a cliche but correlation != causation
causation can only be expressed within your analytical framework. if it not good, then whole world is a chaos and nothing causes anything. causation is pure interpretation.
correlation is there regadless of what you think of it. correlation is interpretation-free.
point being, simplistic correlation != caustion is a mid-wit position.
oh, you mean the interest rates crisis...