LLMs are very useful tools for software development, but focusing on employment does not appear to really dig into if it will automate or augment labor (to use their words). Behaviors are changing not just because of outcomes but because of hype and expectations and b2b sales. You'd expect the initial corporate behaviors to look much the same whether or not LLMs turn into fully-fire-and-forget employee-replacement tools.
Some nits I'd pick along those lines:
>For instance, according to the most recent AI Index Report, AI systems could solve just 4.4% of coding problems on SWE-Bench, a widely used benchmark for software engineering, in 2023, but performance increased to 71.7% in 2024 (Maslej et al., 2025).
Something like this should have the context of SWE-Bench not existing before November, 2023.
Pre-2023 systems were flying blind with regard to what they were going to be tested with. Post-2023 systems have been created in a world where this test exists. Hard to generalize from before/after performance.
> The patterns we observe in the data appear most acutely starting in late 2022, around the time of rapid proliferation of generative AI tools.
This is quite early for "replacement" of software development jobs as by their own prior statement/citation the tools even a year later, when SWE-Bench was introduced, were only hitting that 4.4% task success rate.
It's timing lines up more neatly with the post-COVID-bubble tech industry slowdown. Or with the start of hype about AI productivity vs actual replaced employee productivity.
> like this should have the context of SWE-Bench not existing before November, 2023.
Given the absurdly common mal practice(1) of training LLMs on/for tests i.e. what you could describe as training on the test set any widely used/industry standard test to evaluate LLMs is not really worth half of what it claims it is.
(1): Which is at least half intend, but also to some degree accident due to web scrabbling, model cross training etc. having a high chance to accidentally sneak in test data.
In the end you have to have your own tests to evaluate agent/LLM performance, and worse you have to not make them public out of fare of scientific malpractice turning them worthless. Tbh. that is a pretty shitty situation.
Yes, even if the underlying AI stops advancing today, it will take a while for the economy to digest and adjust to the new systems. Eg a lot of the improvements in usefulness in the last few quarters came from better tooling, not necessarily better models.
But with progress continuing in the models, too, it's an even more complicated affair.
Offshoring was similar - i.e. companies discovered that expensive labor here can be performed inexpensively there while senior laborers/PMs here would perform the overseeing role - and we can look at it how long it took to digest it and adjust to it. While 15-20 years ago it was all the rage, today it is just an established well understood and efficiently utilized, where applicable, practice.
I am inclined to agree, however it wasn't just noticing the difference in wages, but also figuring out how to produce efficiently in the cheaper places and get the goods to rich markets.
Container shipping played a big role in that, and so did modern communication and cheaper flights.
One would expect that if such studies indeed indicate that AI has an effect on early-career workers in AI-exposed occupations, that this would be a global effect. I wonder if there are good comparable non-US studies available.
As a non-US citizen, in my EU country we’re still starving for new programmers.
Poland? Sometimes ago i looked up salaries in Warsaw - it were like $10-$20K/month which as i understand is pretty high by EU standards.
Mostly from US or at least large companies, though. I live in Kraków, some companies offer 10+k USD monthly (Rippling, Atlassian, maybe Google?), for staff level especially. More common range would be 7~10k, i think.
Really? That's crazy. I'm earning a bit over 5k in the Netherlands. Granted, not Amsterdam, but still.
Our guys in St.Petersburg 20 years ago were pulling $3-6K. Granted the life there comes with some risks and inconveniences :)
You're probably working in a domestic company which usually pays less than offshored jobs by a large transnational (and domestics say in Russia were paying significantly less than the offshored). I don't think many companies do significant offshoring into Western Europe though.
before or after taxes ?
Before taxes, always before taxes.
Comparing salaries between countries with vastly different approaches to taxes, health insurance, living cost, hidden side costs etc. is hard and can easily be hugely misleading.
Not even including any subtle things just what "before and after taxes" means can differ. E.g. where I live "after taxes" does for most people not just deduct taxes but also the base health insurance cost and some other things (but to make it more fun, only most but not all people. This means what after tax means can differ between neighbors ).
And then there are so many hidden cost which can influence taxes, e.g. in some areas you practically have to have a car this means that in effect the cost of a car isn't that different from a fixed sum tax, if you consider social standards it even scales with income up to a certain point income like most taxes (and cost x10+ if you can't drive a car for health reasons). On the other hand if you live in a area with decent public transportation and then a car is a Luxus good, but someone has to pay for the public transportation, and if you area isn't overrun by well paying tourists this means you likely pay more tax (but as public transportation tends to scale better you still likely save money especially if you aren't wealthy).
Anyway so comparing after tax in vastly different countries is IMHO a folly. And even before tax is tricky but you have to choose something. I guess another option is "what to frugal living people with reasonable health insurance and rent have left over at the end of the month" is theoretically the better statistic, but just not practical.
It's funny the lengths to which companies will go in order to avoid work.
It's also funny how much they'll spend to save a few pennies
Well...
> The contest between the capitalist and the wage-labourer dates back to the very origin of capital. It raged on throughout the whole manufacturing period. [112] But only since the introduction of machinery has the workman fought against the instrument of labour itself, the material embodiment of capital. He revolts against this particular form of the means of production, as being the material basis of the capitalist mode of production.
> [...]
> The instrument of labour, when it takes the form of a machine, immediately becomes a competitor of the workman himself. [116] The self-expansion of capital by means of machinery is thenceforward directly proportional to the number of the workpeople, whose means of livelihood have been destroyed by that machinery. The whole system of capitalist production is based on the fact that the workman sells his labour-power as a commodity. Division of labour specialises this labour-power, by reducing it to skill in handling a particular tool. So soon as the handling of this tool becomes the work of a machine, then, with the use-value, the exchange-value too, of the workman’s labour-power vanishes; the workman becomes unsaleable, like paper money thrown out of currency by legal enactment.
Karl Marx, The Capital, Book I Chapter 15.
https://www.marxists.org/archive/marx/works/1867-c1/ch15.htm
> Hard to generalize from before/after performance.
While this is true, there are ways to test (open models) on tasks created after the model was released. We see good numbers there as well, so something is generalising there.
> LLMs are very useful tools for software development
That's an opinion many disagree with. As a matter of fact, the only limited study up to date showed that LLMs usage decrease productivity for experienced developers by roughly 19%. Let's reserve opinions and link studies.
https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...
My anecdotal experience, for example, is that LLMs are such a negative drain on both time and quality that one has to be really early in their career to benefit from their usage.
Probably depends on how it’s being used.
I use LLMs every day. They are useful to me (quite useful), but I don’t really use them for coding. I use them as a “fast reference” source, an editor, or as a teacher.
I’ve been at this since 1983, so I’ve weathered a couple of sea changes.
I wouldn't call myself an 'experienced' developer, but I do find LLMs useful for once-off things, where I can't justify the effort to research and implement my own solution. Two recent examples come to mind:
1. Converting exported data into a suitable import format based on a known schema 2. Creating syntax highlighting rules for language not natively support in a Typst report
Both situations didn't have an existing solution, and while the outputs were not exactly correct, they only needed minor adjustments.
Any other situation, I'd generally prefer to learn how to do the thing, since understanding how to do something can sometimes be as important as the result.
People who suck at typing are better off writing by hand as well. I don't need to argue, I'll let history pick a winner.
There are many ways to use an LLM.
Writing code is a bit crazy, maybe writing tedious test case variations.
But asking an LLM questions about a well established domain you're not expert in is a fantastic use case. And very relevant for making software. In practice, most software requires you to understand the domain your aiming to serve.
Not sure why your post was dinged. I have used them for exactly this, and it has been amazingly effective.
> decrease productivity for experienced developers by roughly 19%.
Seems about right when trying to tell an LLM what to code. But flipping the script, letting the LLM tell you what to code, productivity gains seem much greater. Like most programmers will tell you: Writing code isn't the part of software development that is the bottleneck.
I’m 15 years into my career and I write Haskell every day. I’m getting a massive productivity boost from using an LLM.
How do you find the quality of the Haskell code produced by LLM? Also, how do you use the LLM when coding Haskell? Generating single functions or more?
I'm in a similar situation. I write Haskell daily and have been working with Haskell for a bunch of years.
Though I use claude code. The setup is mostly stock, though I do have a hook that feeds the output of `ghciwatch` back into claude directly after editing. I think this helps.
- I find the code quality to be so-so. It is much more into if-then-else than the style is to yolo for my liking. - I don't rely on it for making architectural decisions. We do discuss when I'm unsure though. - I do not use it for critical things such as data migrations. I find that the errors is makes are easy to miss, but not something I do myself. - I let it build "leaves" that are not so sensitive more freely. - If you define the tasks well with types then it works faily well. - cluade is very prone to writing tests that test nothing. Last week it wrote a test that put 3 tuples with strings in a list and checked the length of the list and that none of the strings where empty. A slight overfit on untyped languages :) - In my experience, the uplift from Opus vs Sonnet is much larger when doing Haskell than JS/Python. - It matters a lot if the project is well structured. - I think there is plenty of room to improve with better setup, even without models changing.
I'm stuck in my ways with vim/tmux/ghci etc, so I'm not using some AI IDE. I write stuff into ChatGPT and use the output, copying manually, or writing it myself with inspiration from what I get. I feed it a fair bit of context (like, say, a production module with a load of database queries, and the associated spec module) so that it copies the structure and patterns that I've established.
The quality of the Haskell code is about as good as I would have written myself, though I think it falls for primitive obsession more than I would. Still, I can add those abstractions myself after the fact.
Maybe one of the reasons I'm getting good results is because the LLM effectively has to argue with GHC, and GHC always wins here.
I've found that it's a superpower also for finding logic bugs that I've missed, and for writing SQL queries (which I was never that good at).
Try Claude Code.
Why?
I use similar style as you. neovim with ghci inside, plus hls, and ghciwatch.
Claude code is nice because it is just a separate cli tool that doesn't force you to change editor etc. It can also research things for you, make plans that you can iterate before letting it loose, etc.
Claude is also better than chatgpt at writing haskell in my experience.
LLMs help a lot in doing 'well defined' tasks, and things that you already know you want, and they just accelerate the development of it. You still have to re-write some of it, but they do the boring stuff fast.
They are not great if your tasks are not well defined. Sometimes, they suprise you with great solutions, sometimes they produce mess that just wastes your time and deviates from your mission.
To, me LLMs have been great accelerants when you know what you want, and can define it well. Otherwise, they can waste your time by creating a lot of code slop, that you will have to re-write anyways.
One huge positive sideffect, is that sometimes, when you create a component, (i.e. UI, feature, etc), often you need a setup to test, view controllers, data, which is very boring and annoying / time wasting to deal. LLM can do that for you within seconds (even creating mock data), and since this is mostly test code, it doesn't matter if the code quality is not great, it just matters to get something in the screen to test the real functionality. AI/LLMs have been a huge time savers for this part.
I get the impression that the software scenarios where LLMs do the best on both reliability and time-saving are places where a task was already ripe (or overdue) to be be abstracted away: Turned into a reusable library; as as a default implementation or setting; expressed as a shorter DSL; or a template/generator script.
When it's a problem lots of people banged their head against and wrote posts about similar solutions, that makes for good document-prediction. But maybe we should've just... removed the pain-point.
That's a skill issue. That lone study was observing untrained participants.
It's no surprise to me that devs who are accustomed to working on one thing at a time due to fast feedback loops have not learned to adapt to paralellizing their work (something that has been demonized at agile style organizations) and sit and wait on agents and start watching YouTube instead, as the study found (productivity hits were due to the participants looking at fun non-work stuff instead of attempting to parallelize any work).
The study reflects usage of emergent tools without training, and with regressive training on previous generation sequential processes, so I would expect these results. If there is any merit in coordinating multiple agents on slower feedback work, this study would not find it.
Interesting take. I suggest an alternative take: it's a skill issue if LLMs help a developer.
If the study showed that experienced developers suffered a negative performance impact while using an LLM, maybe where LLMs shine are with junior developers?
Until a new study that shows otherwise comes out, it seems the scientific conclusion is that junior developers, the ones with the skill issues, benefit from using LLMs, while more experienced developers are impacted negatively.
I look forward to any new studies that disprove that, but for now it seems settled. So you were right, might indeed be a skills issue if LLMs help a developer and if they do, it might be the dev is early in their career. Do LLMs help you, out of curiosity?
Imagine if you’d worked for a decade as a dev using Notepad as your code editor (in a world where that was the best editor somehow). You’d developed your whole career in Notepad and knew very well how to work with it
Then, someone did a two week study on the productivity difference between Notepad, vim, emacs, and VSCode. And it turns out that there was lower observed productivity for all of the latter 3, with the smallest reduction seen in VSCode.
Would you conclude that Notepad was the best editor, followed by VSCode and then vim and emacs being the worst editors for programming?
That’s the flaw I see in the methodology of that study. I’m glad they did it, but the amount of “Haha, I knew it all along and if you claim AI helps you at all, it’s just because you sucked all along…” citing of that study is astonishing.
> citing of that study is astonishing and somewhat comical
I would like to see your study, one that's not sponsored by OpenAI or github, that shows LLMs actually improved anything for experienced developers. Crickets.
So, to summarize:
1. An actual study shows that experienced developer's productivity declines 19% when using an LLM.
https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...
2.The recent actual MIT study showing 95% GenAI projects fail to have any tangible results in enterprises:
https://fortune.com/2025/08/18/mit-report-95-percent-generat...
And your source is: 'Trust me bro'. I swear the new LLM fanbase is the same as good ol' scrum: a bunch of fanatic gaslighters.
It's always a "skill issue" , "not doing it right" , "not the proper llm/scrum flavor", or a "flawed study".
When I see the studies, then I might actually listen to the LLM booster crowd, but for now I got studies, what you got? Vibes? Figures.
My argument is limited to “we don’t know and one study with significant limitations with regards to participant adaptation doesn’t settle anything definitively for the long-term”.
Your argument seems to project significantly more certainty and spittle.
The 10% reduction in hiring for young workers is entirely because industry (software, manufacturing) at least in the united states (and probably the world) is contracting and in recession, while the services and government sector has been the main sector growing since a long time now - completely due to economic and geopolitical reasons, nothing to do with AI.
Genuine curiosity: how do you know that software is in a recession? What measures do you use to determine this? And how do you know that the recession is not AI driven? I don't think it is either, but it's more of a feeling; I'm not sure how I would make that argument more grounded.
Well, the best measurement is hiring slowdown and bankruptcies. Bankruptcies in the US are up 13.1%.
Traditionally, you wouldn't look at the release of a productivity tool coinciding with a hiring slowdown and assume that it's automation causing the hiring slowdown, your first instinct would be that the sector is not doing well.
Check sum of free cash flow of major tech co.s. you'll soon find out that the cash needle of industry as a whole is not moving that much.
They are just round tripping the cash that was sitting in their accounts through investments that make their way back through advertising channels or compute channels.
Once you see the bigger picture, you'll realise its all just a Fugazi post covid
Yes. And unfortunately (or luckily for some) the end of the low interest and geopolitical changes aligned with the advent of LLMs.
And what many people don't notice: interest rates are not high.
To me it seems that LLMs are a tool that only increase productivity for given headcount in dimensions that were neglected in the past.
For example, everyone now writes emails with perfect grammar in a fraction of a time. So now the expectation for emails is that they will have perfect grammar.
Or one can build an interactive dashboard to visualize their spreadsheet and make it pleasing. Again the expectation just changed. The bar is higher.
So far I have not seen productivity increase in dimensions with direct sight to revenue. (Of course there is the niche of customer service, translation services etc that already were in the process of being automated)
It's an interesting dilemma, since if I know that an email was written mostly with AI, it feels to me like the author didn't put effort in, and thus I won't put much effort into reading the email.
I had a conversation with my manager about the implications of everyone using AI to write/summarise everything. The end result will most likely be staff getting Copilot to generate a report, then their manager uses Copilot to summarise the report and generate a new report for their manager, ad inifinitum.
Eventually all context is lost, busywork is amplified, and nobody gains anything.
My experience has been interesting. I have been sending super short, mostly dot point emails since before LLMS. Pre ChatGPT I used to cop a bit of shit about it. Now thought, people love it.
> Eventually all context is lost, busywork is amplified
why not fire everyone in between the top-most manager and the actual "worker" doing the work, as the report could be generated with the correct level of summary?
Mostly because there are different depths of reporting required depending who you’re creating said reports for. Often it’s unnecessary bureaucracy, but also often the ones doing the “actual work” don’t have a full understanding of how what they’re working on interacts with other parts of a system. (I mean this broadly, and not just related to software development)
Middle management can sometimes be good at this, because they may actually have the time to step back and take a holistic look at things. It’s not always easy to do that when you’re deep in the weeds with clients, managers, colleagues, or direct reports bugging you about misc things.
Overall I think (or hope) the more useless reporting will die a slow death, but I also think there’ll be a loooooong period of AI slop before we reach the point where everyone says “why are we actually doing this?”
as google seems to do (seen that other HN post today)
https://www.cnbc.com/2025/08/27/google-executive-says-compan...
I'm sad to see this for several reasons because I do not expect or want everyone up use a LLM to converse with me via mail, the whole point is to exchange information, with everyone using a LLM as output and input, now the whole thing becomes a game of telephone.
You do not need to build a spreadsheet visualiser tool there are plenty of options that exist and are free and open source.
I'm not against advances, I'm just really failing to see what problem was in need of solving here.
The only use I can get behind is the translation, which admittedly works relatively well with LLMs in general due to the nature of the work.
> For example, everyone now writes emails with perfect grammar in a fraction of a time.
And absolutely bloody _hideous_ style, if they are using our friends the magic robots to do this.
I don't have time to read paragraphs of AI slop emails. Please keep them short and to the point. No need to send it through an LLM.
Corporations will require everything going through an LLM to meet company standards.
Is 2 years or so big enough sample size for any conclusions? You’re also seeing massive money movements; last quarter larger than consumer spending (!!!) This money isn’t going to junior head counts(labor), it’s going to compute(capital) Also what’s everyone’s balance sheet really saying? Will this money movement to capital rather than labor ACTUALLY pay off ? I think that would take a 10year hind sight to prove no?
I had a team of developers and essentially told them all 'either learn to code with Claude' or you're out. What I found is the more junior developers started 'vibe coding' resulting in a net decrease in performance, where the more senior ones used it to accelerate their speed cautiously and selectively.
My conclusion was senior engineers were better because they were used to managing developers and taking on more managerial tasks building 'LLM Soft Skills' and also frankly fixing mistakes, the junior developers were pressured for speed and had their managers to correct them.
Within 12 months, despite extensive attempts, only the mid level team members remained.
I've got a few buddies over at Microsoft, they've all said something along the lines of "I really hate using copilot. They at least let us use pre-approved models in VSCode, we get most that come out. But all AI metrics are tracked and there are layoffs every quarter. I have kids now man. Strange times. I know you would have quit months ago" and they're right.
Hopefully they have racked up a few million.
https://www.fool.com/investing/2024/11/29/this-magnificent-s...
Any Board which supports management hollowing out future profits by either firing, or not hiring junior staff deserves to have their bonus rescinded.
Think like a forestry investor, not a cash crop next season.
It's going to be extremely interesting to see what the field of software dev will look like in a few years given how few juniors are getting hired recently.
Same as a few years after the dot com crash and after the financial crisis (assuming that this is a real phenomenon; I’m not currently convinced); a critical skills shortage.
(This isn’t unique to IT; this cyclical underinvest-shortage-panic pattern happens in a lot of industries.)
Corporate incentives are usually not pushing in this direction.
Then why are so many corporations reducing staff citing ai?
Stonks go up
Now that bs work has next to no cost, I see a lot more bs work being done, and often on pointless bureaucratic activities involving generating questionnaires and answering them. It's as if the activities add up to a big net zero.
I think mRNA vaccines and green energy are equally transformative economic opportunities. In the US though we are becoming a one trick pony. Instead of investing in all 3, we will prioritize AI because Silicon Valley sucked up to Trump in the recent election.
All to say we could have quite a bit more resilience as an economy, but we decided to sacrifice our leadership in these areas.