Have your company stopped hiring juniors?
Posted by Counter-Business@reddit | ExperiencedDevs | View on Reddit | 93 comments
The company I work for has concluded that juniors do not provide any benefit to the team as the bottleneck is no longer about how fast you can write code, it is how fast you can review the work.
With the improvements of AI we do not need to assign tickets to juniors and review their work. We can build the solution as fast as we can write the ticket. Spending the additional time reviewing the PR of the junior is wasted work and the experienced dev is better off just instructing the AI themselves rather than using an inexperienced junior as a middle man.
My company is laying off all of its juniors and is replacing them with half the number of senior and staff level devs.
So far we are seeing a lot less friction, less tech debt, less time spent on mentoring and reviewing, more trust in the work of our team members. Overall the experience is so much more enjoyable as a dev. Less prod issues, less review needed.
Is anyone else experiencing something similar?
nosayso@reddit
This is certainly the latest "general wisdom" on AI adoption. They tried firing seniors and let juniors write all the code with Claude - that super didn't work, so now the opinion is that your top class senior engineers become productive beyond all reason with Claude and you don't really need juniors anymore.
This is not a sustainable perspective, but it doesn't stop MBAs from doing what they'll do.
21_Wrath@reddit
I am disgusted by these comments ngl. Senior devs owe it to the new gen to hire new devs, just like they were hired during their time. Don’t be so apathetic and self centered people, our profession is going through a great crises and being selfish helps nonone.
lolimouto_enjoyer@reddit
It's not like we're the one making these decisions.
21_Wrath@reddit
The ops lowkey sounds like he's relieved that he doesn't have to work with new devs.
Also you should be influencing these decisions to the best of your abilities.
Stock_Damage2623@reddit
Precisely. And it's the same people who will lament slop and lack of rigor in our industry. It's hypocritical and disgusting
Unfair-Sleep-3022@reddit
Juniors were always much slower than seniors. Your premise is false then.
throwaway_0x90@reddit
In this current AI hype world, it is difficult to justify jr SWEs right now. That's across the board as far as I can see. The industry doesn't need people who are simply experts at leetcode and can reverse a linked list. The industry needs leaders, project managers and system design minded people. The CS degree wasn't teaching people much about these things.
lolimouto_enjoyer@reddit
Yes
stedmangraham@reddit
Without juniors we never get seniors. Just more companies only looking forward to the next quarter.
vxxn@reddit
The price of labor is the feedback loop that will self-correct this over time in the only language businesses understand: money. If we reach the point in future that seniors are scarce relative to business needs the price of labor will rise until it begins to make economic sense to hire juniors again. The problems in the market are twofold:
(1) The salary premium on seniors over juniors is much less than it should be. A seasoned mid-career professional may earn 2x what a fresh grade makes, but their value to the business is far more than 2x the value a junior provides (even before accounting for AI slotting into the economic niche occupied by juniors). Whether you think this means seniors are underpaid or juniors are overpaid (or both), the fact remains that this makes juniors seem like poor value to businesses.
(2) In the current environment, senior engineers are NOT scarce relative to business needs (e.g. every week it seems like there's another big company laying off 10k employees), so there's no need to look beyond seniors. The boomtimes when juniors were most hireable coincided with times when many companies found it extremely challenging to fill vacancies at the level and price that they wanted. It's easier to lower hiring standards than raise budgets (due to corporate beancounters in charge of budgets) which created a pathway for new grades, bootcampers, etc to find their way into their first jobs.
Counter-Business@reddit (OP)
This is true on a macro level but from the individual company level they have no reason to train these juniors if it don’t benefit them now.
Izikiel23@reddit
Tragedy of the commons
Yourdataisunclean@reddit
Beat me to it.
This is the reason I've made a calculated bet to know my shit and stay in tech. Eventually the bills from the lost junior cohort, LLM based cheating, and attempts to get away with lower quality will come due, and there will be an increased market for those who have the skills and knowledge to do senior level work.
Izikiel23@reddit
So, COBOL the revengeance
Yourdataisunclean@reddit
COBOL 2: Talent Shortage Boogaloo.
acommentator@reddit
Only way out I see is to restore mutual loyalty between employer and employee so the employer gets an ROI over time on training juniors.
This can be done, but it is probably not going to happen at most companies with MBAs and dark triad tech bros in charge.
FWIW I arrived at that conclusion before LLMs because I didn't see an ROI in hiring juniors and was bummed out by that.
Famous-Test-4795@reddit
I mean, they do have a reason if they’re not being extremely short-sighted
stedmangraham@reddit
Yes. It’s the same short sighted thinking that has brought industries to ruin. See also stack ranking, outsourcing everything, and union busting.
semiquaver@reddit
Bots will be the seniors in 10 years. Not joking. I don’t see how it could be otherwise.
stedmangraham@reddit
Well if that’s the case then everything is completely fucked and there’s nothing we can do.
I disagree that that is inevitable. LLMs are genuinely astounding in what they can do, but they aren’t people
tnerb253@reddit
And then we have companies that believe people are seniors, spoiler: that's me
Due_Helicopter6084@reddit
Poor guys.
Yourdataisunclean@reddit
The hospital I work for has concluded that interns do not provide any benefit to the team as the bottleneck is no longer about how fast you can diagnose patients, it is how fast you can write the prescriptions.
With the improvements of AI scribes we do not need to assign patients to interns and review their work. We can treat the patients as fast as we can write the prescription. Spending the additional time reviewing the patient record of the intern is wasted work and the experienced doctor is better off just using the AI scribe themselves rather than using an inexperienced intern as a middle man.
My hospital residency is laying off all of its interns and is replacing them with half the number of attending level doctors.
So far we are seeing a lot less time explaining things, less medical questions, less time spent on supervising and reviewing, more trust in the work of our already experienced team members. Overall the experience is so much more enjoyable as a doctor. Less potential patient safety issues, less review needed.
Is anyone else experiencing something similar?
/s
Counter-Business@reddit (OP)
Makes sense tbh sounds like a great idea
ZunoJ@reddit
We stopped a while ago. Not directly because of AI but because the quality of newly grads dropped so significantly that we decided it is not worth to put any work in them
KayLikesWords@reddit
The fact that we aren't training our replacements anymore fills me with white hot rage because mentoring juniors is/was my favourite aspect of this job - but I can't argue with this.
The latest couple of crops of junior developers I've experienced have been fucking useless. I'm finding a lot of young people are coming into the career with zero curiosity about technology. Everything is a black box to them and they are only interested in learning within the bounds of their tools and nothing more. I'd be willing to bet that when most of us were juniors we all had side projects running pretty much all the time. Seeing way less of that.
I don't think there is necessarily anything wrong with just programming for a paycheck but I think this is a much more fulfilling career path for people who actually enjoy making computers do stuff.
Izkata@reddit
We haven't hired anyone for some time now, but from the recent ones compared to the earlier ones, I've been suspicious that they think they'll learn everything they need to know in the classroom before getting a job, and can then just apply it without ever learning anything new once they get a job.
One_Battle7107@reddit
Everyone is coming to their senses that good seniors are needed. The Convex database CEO recently said on a panel that Silicon Valley hiring for experienced engineers has never been tougher.
The real alpha IMO will be in qualifying AI native juniors who are extremely capable. They will act as the forcing function for orgs to adopt AI better compared to seniors.
Blitzkind@reddit
We let our last junior go a couple weeks ago. Not related to AI, he just was a really big under performer.
That being said, we're still looking to hire more last I checked. We have a lot of government contracts so we've been erring on the side of caution when it comes to LLM usage.
localhost8100@reddit
In my 8 years of career across 5 companies. Apart from me being junior in my company, there was only one company hired juniors once. As long as I get work done, they don't give a crap about hiring juniors.
youngggggg@reddit
my company only hires juniors because they can’t afford (or don’t want) to pay for seniors
Counter-Business@reddit (OP)
1 senior is better than 2 juniors IMO
youngggggg@reddit
tell that to my boss haha. They’re pretty good at scouting though I’ll say, most people we’ve hired start making great contributions pretty quickly. Codebase isn’t that big tho which is definitely part of it
Counter-Business@reddit (OP)
How many lines is your code base.
youngggggg@reddit
Don’t know off the top of my head, and it’s spread across many different repos
Counter-Business@reddit (OP)
My company has migrated all projects to a monorepo which seems to help the ai work better across projects. We have 360,000 lines of code and tests.
youngggggg@reddit
Personally I’ve found AI to be pretty good at working across repos as long as they’re all in the same workspace. Monorepo would be more convenient I suppose but putting them all the relevant ones in one workspace for a given task isn’t terribly hard.
sour-kiwi-dude@reddit
Yes my company has stopped hiring juniors. But me personally, I do not think that reviewing juniors code is a waste of time. 15 minutes at the start of my day to review their code is literally nothing. I do it while I sip on my coffee.
It is really unfortunate what this industry is facing at the moment...
Counter-Business@reddit (OP)
Idk if it’s the same at your company but we have each engineer changing hundreds of lines or even thousand lines of code per day. It’s impossible to review all of it. The ownership of the code review is on the engineer themself at my company. Only the plan gets reviewed the code is expected to work as designed.
sour-kiwi-dude@reddit
Yes of course we had this as well. And you're right, it's impossible to manage thousands of lines of changes. But we had a discussion and we made a decision that if you're going to be that guy who's going to make thousands of lines of code changes, you better break it into multiple manageable PRs (good luck with that). Luckily we do have a fair bit of understanding and culture between us, so we have a good manageable process.
KronktheKronk@reddit
Short sighted.
In ten years everyone will be gone and no one will know how anything works
dystopiadattopia@reddit
But do predictive text generators really review code reliably?
ninetofivedev@reddit
Reliably? No. Better than 50% of engineers? Maybe...
vxxn@reddit
Ding ding ding. Most human code reviews are terrible. The bar is on the floor.
another_dudeman@reddit
Fair 😂
Counter-Business@reddit (OP)
LGTM
iagovar@reddit
That's my boy
dystopiadattopia@reddit
🤣
semiquaver@reddit
Are you talking about the LLMs or the senior engineers? 🥁
Counter-Business@reddit (OP)
If you front load the review to the planning stage, and you make sure the business logic is solid, this corrects the majority of the bugs. The problem is that juniors really struggle with business logic and often fail at the planning phase, leading to failures cascading downstream from the root (the plan) into the entire implementation. Seniors are much better at utilizing the AI because they know when and how to critique the plan due to the experience.
dystopiadattopia@reddit
Yeah, it seems like almost everyone says the most appropriate people to use AI are those who don't need it.
franktronix@reddit
Seniors don’t need it is not a strong argument. Seniors don’t need anything besides vi for editing code either right? AI is a tool that had huge benefit when used right.
Counter-Business@reddit (OP)
If you have to use it as a crutch then you trust it too much which is really bad because an unchecked AI will go rouge and make lots of tech debt.
dystopiadattopia@reddit
Unfortunately way too many people use it as a crutch. They think whatever it spits out is the best possible answer, when it rarely is. That's what worries me most about AI - the people who use it.
Counter-Business@reddit (OP)
Yeah I totally agree with your sentiment. The only thing I slightly push back on is that seniors don’t need it. I believe if in the right hands, (not the wrong hands) it can be a net positive. But in the wrong hands (inexperienced) then it wrecks havoc on the code base.
upsidedownshaggy@reddit
I'm going to be honest if your juniors don't understand/struggle with business logic that's a failure on your team's leadership for A) being bad at teach, or B) being bad at weeding out people who don't care enough to learn if you have the materials available.
Vlookup_reddit@reddit
it doesn't. but management doesn't care.
donniedarko5555@reddit
Also if your PR gets merged and it breaks production they come to you saying "why did your code fail".
PR's were always a last step in a chain, primary responsibility is on the developer. As an org I think AI or not isn't the important question. Its how do you let developers know they are accountable for the code they commit
ninetofivedev@reddit
It's.... far from the last step in the chain. My change doesn't break prod. It typically breaks in dev.
Sure, maybe we can shift this left to catching this before the PR is even merged. That would be ideal, but most companies are still on very archaic branching strategies and deployment processes, but they're still not going straight to prod.
Vlookup_reddit@reddit
i mean at this point if there is a credit you have to thank ai like 10 times, but if there is an "accountability", you bet your ass that it's solely on you.
at least that's how 2026 q1 feels like, unsure about q2 tho
sarhoshamiral@reddit
Seems like sweet spot for companies is mid-senior people.
But question is what happens when they retire?
Franks2000inchTV@reddit
I think looking at this as a cost-cutting measure in general is short-sighted.
Lets assume that you can obtain some measurable performance increase with AI. So an agent-developer pairing can put out n-times more work than a developer alone.
And lets say there are two competitors who each have 50% of the market share in a given market.
Competitor A reduce their workforce to 1/n and keeps 100% productivity.
Competitor B keeps the same headcount and costs and increase to n * 100% productivity.
Who's going to be doing better in 5 years?
jaymangan@reddit
You’re not accounting for the new bottleneck. Don’t think of it as AI for a second, and just as new tools and dev processes.
When n is 1.1, or even 1.5, it looks amazing and you’re seeing the old bottleneck of delivering tickets shrink by 9-33%. But what if n is 10? Or 100?
If we assume perfect executive decision making (hypothetically), then in the past product was almost always ahead of development. At 10-100x, it isn’t. Now product needs to find order of magnitude multipliers to feed tickets to the dev funnel.
What happens when product does? You’re delivering so much so fast that you strain executive decision making to the point of requiring them to be psychic, and you need to find a way for your users to stay up-to-date, and your customer care teams have to field more tickets (they better find tool adjustments as well).
My point is that old methods are no longer competitive without being re-envisioned. The same problems exist, but past processes assumed a scale for each phase that is now out of proportion. They need to grow together, and there’s also a limit to what a company and its audience can do, so there can be a racing towards the asymptotic limit as well.
My unsubstantiated working belief is that the industry will move toward smaller teams that do “agile experimenting” instead of “agile development”. (I know XP and other processes allow experimenting. But they are steeped in development. I mean the phrases and discussions and rituals will be about experiments, and the development will just be an assumed detail.)
And maybe some niches like we used to have DBAs, but for DX. Job is to stay ahead of the curve and ease the rest of the team’s tool churn and onboarding.
barley_wine@reddit
How many big business decisions in a publicly traded company are anything but short sighted, they just care if it raises stock prices today and reducing your employee count does usually result in higher prices.
upsidedownshaggy@reddit
If I had to guess it'll be a similar situation to what happened with trade jobs here in the US. You're gunna see a bunch of annoying business owners complaining about how no one wants to work for stagnate low wages before retiring and the new owners are going to have to pay out the nose just to attract basic talent.
WholeBet2788@reddit
Funilly enough no one cares because thats 20+ years from now and if ai will continue with this tempo amount of people needed will dramatically lower. Also how many companies even survive 20+ years before bankcrupting or being merged. Realistically the companies today not hiring juniors wont be there in 20+ years to pay the due.
slackwaresupport@reddit
they need some real world test before hire. ive had jr's who after 2 yrs ask day 1 questions.
Careful_Ad_9077@reddit
Why put the effort into milking the cow when you can just kill the cow and sell the meat?
bigmoneyclab@reddit
You are the one getting laid off mate
13--12@reddit
Never really hired juniors, only interns. And we never stopped hiring interns.
Dry_Bird1790@reddit
I think juniors just call themselves seniors now
_5er_@reddit
Juniors were never hired as a super productive workers. They are mainly investment for the future.
This whole shennanigans about "we don't need juniors because AI" is the worst argument ever.
And at the end of the day, juniors can also use AI. You just need to set very strict guardrails for them.
Devboe@reddit
Yes, we haven’t hired a single junior or even mid level in the past 4 years. Only senior and above
JL14Salvador@reddit
As great as you make that all sound right now it’s also not great for the future of the industry if no juniors are ver being hired. How will we mov e forward with the next wave of senior engineers for the future if we can’t foster juniors to get them there. Not a concern for senior engineers today but companies will feel the pain in the next few decades when there are no more senior engineers to hire.
c0ventry@reddit
We just hired one a few months back and he is fantastic.
szansky@reddit
so hire me as well xD
ExperiencedDevs-ModTeam@reddit
Rule 1: Do not participate unless experienced
If you have less than 3 years of experience as a developer, do not make a post, nor participate in comments threads except for the weekly “Ask Experienced Devs” auto-thread.
Adept_Carpet@reddit
I feel juniors have a very valuable and viable role though they are needed in smaller in the past.
It used to be, they had this enthusiasm and fresh ideas and lack of cynicism. They could really dive into a project and believe in what they were building, but the weakness was the code.
Well, now everyone on earth is somewhere between mid-level and superhuman at coding depending on stack/task/context etc. You get their energy and perspective but you don't have to put up with the lousy code anymore.
A team without a junior is incomplete, except in special cases.
Lfaruqui@reddit
From what I can see….it’s only senior and up at my office
Mortimer452@reddit
My company did this well over a year ago. Seven devs, three senior four junior, laid off all the juniors go and went all-in on AI tools.
Since then, changes are shipping faster, getting way more work done. When issues arise the choice between "quick dirty fix" and "long-term correct fix" is no longer a struggle since they both take about the same amount of time.
Honestly it sucks, I'm not sure WTF the industry is going to do in 5-10 years when the majority of "senior" devs are basically just prompt engineers with almost no ability to actually read code, but here we are
Altruistic-Bat-9070@reddit
In 10 - 20 years senior devs are going to be able to write their own pay check with no new ones being trained
ninetofivedev@reddit
So... Another theory might be that we see shrinkage in the SWE job market, basically for the first time ever.
In which case, this might not be true. We're still going to have a longer term funnel problem. A baby boomer like event, but we see that all the time.
UnbeliebteMeinung@reddit
Sounds good for me.
DollarsInCents@reddit
My company stopped hiring all levels 2 yrs ago and pivoted to Pune. The new hire introduction slack has been dry af
0xSEGFAULT@reddit
My company has switched to “near shore” hiring in the Dominican Republic.
:|
bunkkin@reddit
They won't hire juniors and they won't do promotions but they will hire senior roles directly.
So that's fun
ninetofivedev@reddit
Yes. And long before AI. Honestly I haven't worked at a company that has hired more than 1-2 juniors over 1-2 year periods since 2019.
driftking428@reddit
No. We have interns, juniors, mid, and seniors all being hired.
VelumLucis@reddit
Yup. Late stage startup, \~1000 employees and 400ish engineers, we stopped hiring juniors 6-8 months ago.
Lechnerin@reddit
Ye . Like 2 years ago we still hired 2 . Now they are mid
ham_plane@reddit
I've been at my current company for 2 years now, and we just started hiring juniors for the first time, about 1 month ago.
Empty_Expressionless@reddit
Yeah I'm interviewing now. Our requirements are generated telephone style customer to analyst to project manager to Dev. By the time a ticket gets to us it's so far been impossible for any agent to accurately reverse and infer what the customer was actually asking for and what their expectations are. We just don't have the context written down anywhere
nonya102@reddit
We stopped hiring altogether and got no ai tools that are any good!