Ask HN: What will tech employment look like in 10 years?

I know you’re trying to see where the puck is headed but I think there is a lot of work between where we are and where LLMs replace 4 Junior devs.

The workflows we have are not quite right for it. Coding has always been 10% coding and 90% debugging but I think the rate at which we generate the 10% will grow exponentially.

This means that the debugging has to grow. We will generate errors at an unprecedented rate.

LLMs trained on previous errors and methods won’t catch them. They’ll be more complicated and spread out over the code.

We need new tools to visualize the code and track errors. I think what it means to be a programmer will change. More testing, thinking and less klocs.

I was going to comment that there will be a need for many IT professionals to fix problems produced by AI. Your comment painted a much clearer picture.

I am not a super coder but I am quite conscious what I write.

I copy pasted a part of my code into Claude.ai to help me add a few lines. It completed that part and also rewritten another few lines. I asked why it did that, explanation was that I missed a edge case.

LLMs are good for writing boring parts of code and also helpful for catching bugs.

You have to be careful because they will sometimes make mistakes they do that. I’ve had Claude decide to “simplify” my code by removing a fix that took hours to chase down.

Was your fix commented? If not, a human developer refactoring your code could have done the same. I’ve seen it happen more than once.

> a lot of work between where we are and where LLMs replace 4 Junior devs

The projection from current SOTA is that coding LLMs replace 1 senior dev + 5 junior devs with 1 senior dev + LLM.

Ergo, there will be fewer junior dev positions and senior dev salaries can capture some of the additional value they deliver.

Whether that's offset by the increase in senior supply as junior -> senior conversions happen...?

The two areas that will likely expand demand to absorb those junior devs are:

(1) Integration engineering (i.e. poorly-documented interface to poorly-documented interface)

(2) Testing (because if kloc are cheap, there will be more money to validate -- and it will make people feel safer to have human-in-the-loop before sign-off / production).

Coding LLMs are always going to be better at things well-represented in their training sets. Which is to say, the most popular languages, API styles, apps as of todayish.

So, other things will be the best use of human time.

> The projection from current SOTA is that coding LLMs replace 1 senior dev + 5 junior devs with 1 senior dev + LLM.

Thanks - I needed a good laugh.

You think junior devs, as the positions exist today, are going to be sticky in the face of coding LLMs?

How else are you going to get senior+ level engineers going forward?

Ha! That's the rub: we won't. :(

In the same way that we don't get the volume of true hackers that the 70s-90s produced, because of the locked-down computing experiences most kids were introduced to from 00+.

Yes. But why would a company today hire a junior dev? If I have limited budget, why hire a junior dev instead of poaching a mid level developer from somewhere else?

A junior dev does negative work. And with corp dev/enterprise dev the difference in price between a junior and a “senior” dev is only about $70K. The gap is even smaller with remote work even if you hire in the US. There are plenty of developers in low cost of living areas who would be glad to work for relative peanuts.

As far as where will the next generation of developers come from? That’s a collection action problem

Sometimes you want them "in 2 years from now" so you 'buy' them now. Also sometimes you want someone to do 3 out of 10 tasks of the job for a couple of years, and then get them to do another 4 out of 10, an ultimately get them to do 10/10 in 6-7 years from now.

I remember those years, I hope they are still around.

In 2 years from now, you’re going to want to promote them but your HR department is not going to give them the raise that would keep them there so they are going to find another job. They are going to watch new people come in in two years making market rates - salary compression.

The average tenure of a developer is 2-3 years. Almost no one stays at a company for 6-7 years

As a manager, the chance of you staying at a job for 6-7 years are slim and staying in the same position is even slimmer. Besides, you are going to be measured by your quarterly or at most yearly “OP1 goals” or equivalent

Smart juniors shouldn't be juniors long. I think another angle to think about this is Junior devs that can leverage AI to learn have a much shorter path to Senior than current Seniors had to take (aka the hard way). If as a Junior you know of things to look out for but don't know the implementation details, AI can help you.

For really obscure issues or bugs, I had to scroll through seemingly random StackOverflow results (or Experts Exchange, ha) looking for exact issues that I had, weeding through answers that weren't relevant, things based on older versions, etc. Now you give the context to AI tech stack, data, and your source code, explain the problem with the error message, and ask what might be happening and that shaves a few hours off the old way of trial and error. Even if the 1st answer isn't totally right, a few follow on questions will help narrow it down.

The hard way makes for better quality.

I can have 18% BMI with ozempic, or I can have 18% BMI running and lifting weights. "We are not the same" to quote the meme [0].

The LLM-bros can do LLM-bro stuff. The hardened Seniors that have seen/done it all are the ones you need on a crisis.

[0]: https://i.imgflip.com/5yhw7e.jpg

The tradeoff is that searching for answers develops people's ability to do research in general. It's an essential skill to be able to skim though material quickly and decide if it's relevant or not. it's in the same way that calculators save labor on large calculations, but giving up all of one's mathematical intuition in exchange for a better calculator would be decrease in net performance.

Massively kicking out junior devs is ignorant and short sighted. Critical IT areas are going to implode understaffed by retiring seniors and freefall quality of AI slop retrained on more AI slop.

As GenX/Millenial senior, I am not going to stay long here, no matter what money I am offered. If cornered I would rather sell organic produce, honey, wood work, be handyman to my local community than spending any more days on debugging fruits of MBA-driven AI lunacy.

The day to day of a software engineering job has changed and will change more depending on how advanced llms get (and all bets are off if we do get to AGI). However, I don't really think you will see a massive drop-off in employment for two reasons. First, competition is going away. If your developers are x times more productive, so are your competitors developers. I can see a bit less need in non-product companies if more domain experts are able to build their own tools, or generalized agentic tools eliminate some of the need for custom tooling.

However, I think there is a second thing that is often overlooked here. It seems the angle is always 'oh companies can replace developers' but no one seems to consider that developers can replace companies. I think you are going to see small teams of very skilled people replace able to make amazing products. There are limits to what llm's can do on their own but a skilled engineer who masters the tooling and can unblock the models and knows the techniques to keep the models productive as the codebase grows will be able to produce amazing things far beyond what they could ever produce before. I think you are going to see an explosion of new, smaller companies. I think if you are an engineer you shouldn't be gloomy, you should be excited.

I think you're even forgetting where the real strength of AI will always lie. AI can reliably make a lot of dumb, simple decisions. AI will massively expand what we can automate, and we can hope that this will bring back, things like textile manufacturing, orchards, large scale agriculture (even where "complex", like harvesting), delivery services (e.g. within companies) ... and all these new firms will suddenly find they require programmers, developers, embedded development, electrical engineering, ...

A LOT of new to be created firms will suddenly find they require programmers, developers, etc. Hence a lot of jobs.

We'll need a small level of improvement in the current crop of robots available, but not much.

I think the developer market is about to expand massively if tariffs are spreading (because that will force a lot of countries to move a lot of production back onshore, as goods are prevented or obstructed on the borders).

Of course, either I'm totally wrong or people haven't yet really realized they can do this, as it's not happening on the level I'm predicting ... yet.

Yes, you rarely hear specifics like this as to where new jobs will come from. Hear, hear!

How did manufacturing adjust to the rise of 3d printing and the new world where 90% of consumer goods were printed at home, replicator-style?

How did the finance world adjust to a world where financial transactions were automated on the blockchain using smart contracts?

How have cities adapted to the massive migration of in-person experiences to the metaverse — and so soon after they rebuilt all their physical infrastructure around the revolutionary personal transportation system known as the Segway[0]?

[0] none other than Steve Jobs predicted that cities would be designed around the Segway https://www.theguardian.com/world/2001/dec/04/engineering.hi...

But it's different this time! Just like it is every time, until it isn't.

(Not to say that revolutionary world-changing progress is not _possible_, but for every Industrial Revolution there are quite a lot of Segways and metaverses.)

It seems a better comparison would be how to programing change with the advent of higher level languages and frameworks? These did have major impacts.

I disagree that LLMs are comparable to a higher level language or framework since the fundamental output it still at the same level of abstraction.

This isn't C to Python, it's Python to more Python faster.

In 10 years, I suspect we will have more off-shoring, extremely few entry/mid jobs on-shore, and an increasing push towards being tech lead like guides. I think companies will use off-shore labor and LLMs to do most of the grunt work while we have senior+ engineers focused on guiding those people/agents.

It's just one more step in the multi-generational trend of eliminating/outsourcing lower skilled jobs and increasing the barriers to entry for the remaining jobs. I'm not optimistic.

Of all the speculation in this thread, I unfortunately suspect this is the likeliest outcome.

Hailing from a popular outsourcing destination I think the GP missed the mark somewhat.

Outsourcing isn't just grunt work. It's almost all work, but cheaper and with less control over the outcome. There will be no "guiding people/LLMs".

You have talented people all over the globe and to date what the best of the best were doing is move to the US. Those were the people guiding/managing outsourced teams, not Americans. You see this now with the "Indians hiring Indians" phenomenon.

Trump put the brakes on that, so currently the only thing that's certain is uncertainty.

My take is that there will be an attempt to replace at least the bottom end of outsourcing with LLMs, so more pressure on offshore labour to perform and, over time, fewer positions overall, as junior roles have already evaporated around here and eventually that will translate to fewer seniors.

"Outsourcing isn't just grunt work. It's almost all work, but cheaper and with less control over the outcome. There will be no "guiding people/LLMs"."

It might not be only grunt work, but it is mostly grunt work. That lack of control has led to our org only outsourcing grunt work and low risk projects because of the higher failure rates we've seen. We are starting to use on-shore senior+ and TLs to oversee the work the off-shore resources are doing to help control the outcomes more.

That's just you and your organization, which is why you're off the mark. Others outsource all kinds of work because talent is not exclusive to the US, but acquiring it might be too difficult.

Also FAANG and the like perhaps didn't outsource, but definitely offshored much of their operations.

I have friends in other companies saying they're doing basically the same thing. I guess unless one of use has a Gartner subscription, we won't know who is more right.

How do you put tariffs on code? heh

A tax on all import statements.

that's dependabot

When AI is cheap, why use off-shore labor at all? AI replaces it. Why not use the best available talent to oversee and guarantee the work of coding agents? To define requirements and translate them into maintainable code? This is the trend.

Outsourcing/offshoring has been happening for the last 20+ years though. If anything, LLMs would remove the need offshore things in the first place.

The idea that you could just learn python and make an upper 5% salary in the first place was absolutely wild in the first place. I've been screaming at anyone who would listen for the last 2 decades that programming is a skilled trade and it's practitioners should be unionizing. This is more true than it ever was. It won't happen, the HN cadre will continue to turn up its nose at the idea until they are slowly incinerated by the roaring flame of capital exploitation.

Everyone is talking about AI this and AI that. I think the more pressing job market pressure is from there being way too many graduates due to increased interest in the field from the general population. I TA at a university and I see the current crop of CS students and the demographics have changed a lot.

Have you noticed the base level knowledge change? I talked with a second year CS student a while back who didn't know how to ping another host, and it really made me wonder if new students are choosing CS because it seems cool (versus _back in my day_ when we were all geeks who took CS because we lived and breathed it all day in life anyway).

I got into CS because I liked games. I wouldn't know how to ping a host by year 2, in fact, I hated CS and only finished it because my father insisted.

Wasn't until I got my first internship by year 4 that I started to work/enjoy it, and have had a very successful career so far (20y in)

Same as it did it 10 years ago.

Generalists will be undervalued, but will always be able to find a job.

There will be pockets of extraordinary cruelty and pockets of extraordinary grace.

Where you find yourself on the cruelty/grace/specialist/generalist surface will still be up to you, but also, just as it has been for thousands of years, also up to fate.

Do your best, control the things you can, and accept the things you can't. Remember that work is just work, not living.

Hang in there, it's only 10 years more to get there...

But there's no living if you can't find work :P

It'd suck a bit should you be automated out of the thing you've spend a few years studying if you're a fresh grad, not to speak of anyone with more experience. While I guess one shouldn't be complaining, no work bring shame, going back to cashier, or warehouse worked would be a bit of a letdown.

A generalist cannot, by definition, be automated out of their specialty.

If the machines zig, you zag.

I think over the next decade many countries are going to flex their autonomy and replace "Big Tech" with their own tech so there will certainly be opportunity in the EU, in UK/AU/NZ/CA, in Latin America. The whole supply chain from embedded software to drivers and operating systems and user-facing software and platforms is going to be under more scrutiny than ever.

One thing that is constant is the billions of people using the internet and their buying power, so I think there will also be tremendous opportunity for people able to release good software while mid+large+huge tech companies focus on eliminating the expensive tech salaries that built and sustain their fortunes. Their software is going to become shittier until LLMs get much better.

It will be glorious! LLMs Will replace most managers and all executives. Developers will use their corporate AI to answer all questions from the outside world, allowing them to focus entirely on writing code. No need to talk to other humans, aside from other developers. Coders will develop lasting personal relationships with their IDEs. AI enhanced design tools will provide pixel perfect designs that users always love. Glorious.

I’ve been working with a team of so called ‘junior’ devs and coding is the least of the problems. Design and architecture is by far the most difficult thing for people to understand and get a hang of. A staff engineer won’t replace 2-4 junior devs with LLM. He will focus on the design and architecture and then get 2-4 engineers to execute it with LLMs. The two might sound the same but it’s not. And that’s the difference between understanding software engineering and coding. LLMs make coding easier, not software engineering.

> A staff engineer won’t replace 2-4 junior devs with LLM. He will focus on the design and architecture and then get 2-4 engineers to execute it with LLMs.

That's already tried (not with LLMs). Especially with CASE tools, UML, and before those, the waterfall process. Building a software is like writing. Only once it's done, you can be really sure you understand the subject. The skill part is to make sure someone else can.

Doing design and architecture is like drafting plans before a war campaign. It helps with planning, but it does not help in winning the battle. There are pocket of complexity that no design can touch, because you would just be writing the software already.

Try this experiment. Install a few apps, like Time Doctor and Wakatime to track where you spend most of your time while on the computer for work purposes.

After a week, take note of how many hours you spend actually writing code.

I won’t share my exact count, but it’s shockingly low in relation to all of the hours spent working.

My bottleneck to more productive output is 100% not “unable to write more code faster”. It’s actually people. Other people.

There are so many directions it could go in. I imagine we'll be revisiting trends from the past - LLM generated apps will be everywhere and behind the curtain, they'll be a lot like a VB6 app from the 90s - System Analysts will be back in vogue because you've got to feed the LLMs something.

10 years after that will be interesting. Can you imagine a $100m business running on dozens of apps generated by various LLMs. Are management going to sign off a rebuild from an LLM or are they going to get a team in to do it from scratch and consolidate the systems.

Agile principles will be back, "Individuals and interactions over processes and tools" will be a popular line.

I was thinking about LLMs, productivity, the 'human tax', and other parameters from a different view.. "What if" (I am not counting the costs yet/at this point)(numbers come from my head - no supporting documentation for them):

  existing team produces 200
  2/10ths of the team + LLM produces 200
  existing team + LLM produces 4000

Of course that varies on the area/dept/function/etc. If I was a businessman, I would seek to see how with the same costs I can explode my productivity (quantity _and_ quality), keep team sizes the same and well fed, and let them take the business from +200 to +4000.

Now, I also have (constantly) in mind Orwell's "in front of your nose" [0], so some-one/some-company may sacrifice (as they usually do) the long term for the short term (shareholders must get their dividend _now_ or else...) and in which case, 'never let a crisis go to waste' let's all fire 20% of our staff, destroy benefits/salaries, and hire back everyone at 20%-30% less. Yes our companies will have less profit in the short term, but the cost saving (2y no pay, and afterwards 30% cheaper labour) would be great.

[0]: https://www.orwellfoundation.com/the-orwell-foundation/orwel...

My bet is that engineers will have to go down, closer to the silicon level. The higher-level stuff will become more and more accessible to the non-tech consumer, so building apps and stuff will no longer require tech-savvy people, everyone will be able to make their own app.

i'm a low-level developer but I disagree, low-level C or chip design or firmware development are actually a lot simpler than (complex) app development, AI can take over low-level way more easily than upper level sw devel.

Tech employment in 10 years is going to favor candidates who know how to get the job done with the fewest moving pieces.

The current era of making the most complicated fucking thing possible is clearly nearing its conclusion. The writing is on the wall with the now weekly conversations about vendor and dependency management crises.

One might say to use something with more batteries included or whatever. I'd go 5 steps further - have you solved the problem outside the computer yet? Do you understand the domain model? Do you understand why the customer wants to pay for any of this bullshit in the first place?

I think that's approximately what it's going to look like. Excel, bash and powershell replacing kubernetes clusters.

I want to agree because I think you and I have similar values. But experience has shown that people tend to add things to obscure problems, never simplify. So while you and I would like to kill those kubernetes clusters with a script, the company selling solutions to CTOs will introduce 'Kubernetes Prime' or some bullshit and talk about how "they heard you liked containers, so they put containers in your containers, now you can virtualize your virtual lives."

At some point, the CEO is going to ask the CTO why the product hasn't been delivered. If this doesn't happen, the board will eventually intervene and install a new Cxx suite.

The financial pressure will eventually correct these issues. The cost of borrowing is non-zero now. Investors are going to want to see happier customers much sooner than before. Hobbling along with a handful of marginally happy customers for a decade isn't going to cut it anymore.

The most likely corrective measure is that the company goes away because they couldn't course correct in time. This is often the tragic consequence of tech egos (often the CTO) refusing to be sublimated in service of the actual business needs.

Kubernetes Prime that's pretty funny because it will probably happen. Reminds me of the South Park episode where they go into the future, and all the companies have merged and either have Prime or Plus or Max in their name, such as the restaurant they eat at is called Denny's Applebees Max.

I think there will be a crash of any sort of junior positions and the employment numbers will fall down. There will be need for experienced, intelligent developers, but only highly below average.

Average devs will be no longer needed as all they can do a senior will be able to do in the same time with LLM.

I think we'll go back to the system as in the Mythical Man-Month - one lead developer, one below them to do less important tasks, and a few domain experts not related to programming. Ironically I think good front-end developers may remain useful as UI/UX experts.

> Ironically I think good front-end developers may remain useful as UI/UX experts.

It's not like they've done a great job so far. There was a time companies spent time and money on testing their software on the people who would use it - nowadays we have "experts" who know it all from school.

I can't say anything on how work will be organised, but there's a thought I had the other day:

The rate of adoption of new APIs will slow down considerably.

LLMs only really know what they're taught and when a new API comes out, the body of learning material is necessarily small. People relying on LLMs to do their jobs will be hesitant to code new things by hand when a LLM can do the same using older APIs much faster.

Who is going to do it then? Well, someone has to or else the API in question won't see widespread adoption.

I predict the same thing, that publishing new APIs will be rare, but for a different reason.

People will stop publishing any new knowledge or method to keep it from the LLMs and to keep people from exploiting new ideas before they can.

I think we’ll see a lot more software. Lots of non tech people will increasingly have the ability to create custom software tools and prototypes. They’ll share and remix these. Some of them will be fully productized, at which point professional software engineers will be called in to help untangle the LLM spaghetti and create proper applications. So I think there will be more software work in consultancy.

I think we’ll see advancements in robotics and more hires there.

And I think there will be more jobs around the LLM ecosystem — progress on foundational models, inference optimizations, on prem migrations, networks of agents, AI more deeply integrated with existing sw.

Overall I think there will be more jobs in observability, security, and infrastructure.

I agree there will be fewer junior positions. I’ve written about some of these ideas before including a deskilling for new practitioners https://matthewbilyeu.com/blog/2025-03-08/ai

Here are my predictions for 10 years out along with upskilling recommendations: https://open.substack.com/pub/drewhoskins/p/the-most-promisi... I agree with many of your points, would just generalize a couple of them. For example, I think there are many pieces of software that need to be complex and excellent, not just those close to hardware. And niche apps are a good (but not only) example of where LLMs will struggle to solo-code because they lack product context. Other key questions for a project: how much training data does the LLM have access to? (Innovative things generally have much less.) How much ownership skill is required to build it?

I think it would be a good idea to start by reading Asbhy and Wiener on the foundations of cybernetics. They knew the limits of using computers to automate stuff. They wrote their books in the 50s. Since then we have tried and failed many times to automate all the things with computers.

In 10 years, resource, energy limits, and wealth inequality will become increasingly clear to everyone, while the climate continues to warm at an accelerated rate. Social instability might preclude the existence of software engineering, aside from a tiny minority supporting the lavish lifestyles of the ultra-wealthy.

Gen Z and the generation afterwards, having had parts of their psychology destroyed by Tik Tok, ready access to AI for cheating school and such, will lack the attention span to properly learn time intensive skills such as coding and first principles engineering. Knowledge of complex maths and theorems will be lost and there will be generally no interest in things such as functional programming, proofs, math, etc.

Initially there will be a huge employment downturn as organisations follow the pattern of hiring a Gen Y expert and supplementing them with Gen Z juniors using increasingly good AIs.

After a while, the Gen Y seniors will retire and/or die and it will be impossible to source a senior developer with actual knowledge. Wages will skyrocket for a few Gen Z experts, but in general there will be a shortage and the entire industry will eventually reboot, although I am stuck to speculate how exactly.

At some point, we will lose the ability to quickly make advancement in AI because engineers will stop understanding how they work due to a lack of deep understanding in mathematics etc. and advancement will drop by 100x or 1000x.

Due to people becoming very lazy on average, capitalism will become ineffective and most governments will veer towards centralised services, or central planning/communism to cope with the lazy populous. A few people with work ethic approximate to today’s founder work ethic, or deep skills, will be worshipped.

No one knows or even has the slightest idea. Predicting the future is a fool’s errand.

When AI is cheap, why use juniors, regular software engineers or off-shore labor at all? AI replaces these things. Why not use the best available talent to oversee and guarantee the work of coding agents? To define requirements and translate them into maintainable code? This is the trend. It takes mastery to produce good, maintainable code from vague requirements using coding agents. This seems like value add a human will continue to provide.

I also see the blending of engineers and data scientists with product managers.

The pattern you've said makes sense. The amount of full stack engineering roles will reduce with LLMs and agents able to complete these tasks with the required context.

Computer Engineering domain will continue to grow. This will be due to 1) need for better GPUs and CPUs 2) need for data centre infrastructure engineers

LLM ecosystem will also continue to grow along with Machine learning ecosystem. Regression based models cannot be replaced by LLMs due to inherently not being reliable in producing consistent output for structured problems

There's a lot of hope in the other comments that senior engineers will survive. And who knows, maybe something will occur to slow AI development. Or maybe we'll hit some hard limit we can't anticipate. Maybe society will even lurch left and act to protect jobs. But if none of these things happen, there will be a day when senior engineers find themselves unemployable, like pretty much every other worker.

Single function hardware machines that pass new weights and biases based upon the latest state and the prompt.

Tech work will look more like a data center tech. Demand for programmer jobs will shrink to only what the hardware manufacturers need.

There will be no tech employment in 10 years in the same way that there were no horses in transportation employment by the 1940s.

For tech writers, it'll mean becoming editors more than writers. Editors of LLM generated stuff. For developers, I guess the evolution will be similar: more architectural thinking, more code reviews, less typing boilerplate.

It's difficult to imagine what it will be like, but it will be radically different. There will always be a need for a programmer to be able to review code and make sure things are going in the "right" direction. One can speculate if this one human + AI will replace 10, 100, or 1,000 human workers.

Recently, there was an analysis of when we might get to Post-Labor Economics:

> 2025 to 2030: Collapse of knowledge work. The "KVM Rule" applies: any job you can do entirely with a keyboard, video, and mouse will be fully replaced.

https://x.com/daveshapi/status/1916188978727784847?t=9YNl90V...

AI + human replacing 1000 humans? This seems like the 9 women can have a baby in 1 month fallacy all over again.

It will be interesting to review this thread in 10 years time.

> There will always be a need for a programmer to be able to review code and make sure things are going in the "right" direction.

I'm curious, what makes you think this?

Maybe companies will just want someone to hold to account.

Did you see the recent case of the airline’s customer service LLM that promised someone a refund due to a relative dying? And the courts forced the airline to honour it despite it not being accurate (turns out LLMs have more humanity than a corporation).

Putting your company in the hands of an LLM workforce could save a few dollars, but it can cost a lot too.

>> 2025 to 2030: Collapse of knowledge work. The "KVM Rule" applies: any job you can do entirely with a keyboard, video, and mouse will be fully replaced.

Even that seems too fast. There will surely be more innovation, some unanticipated, but not that knowledge work will be all gone in five years from now.

In 2030 people will be using LLMs to predict that 2030-2035 will be the end of the knowledge work. It won't ever end. Most jobs are already not needed, more automation will create more useless jobs.

"a single staff engineer with experience and a trusty LLM is similarly productive as a team of 2-4 junior engineers"

I think we are SEVERELY underestimating the amount of slop that is going to come from this.

Depends on how sloppy 2-4 junior devs are and thr overhead of managing and developing them. I think the comparison could be similar

A lot more bad logic will get by the smell test when you have 1 staff member vibe-coding. And just wait until he is held to that level of zero-test, lowest hanging fruit productivity as being "normal".

That is a good point that wasn't in the original description. I do think test engineers will become more important to ensure the generated code actually fits the requirements. We might even see the role evolving towards more of a business analyst role using LLMs to generate the tests and then adjusting/validating them based on the requirements.

The thing about a junior dev is that often times you can convince them to start/stop doing things a certain way.

There's already skill issues in the field even with senior people and without AI. It's why we test people to witch hunt for those imposters. Lots of talk about how one bad hire can ruin the whole team or something. Even competent developers stretching across multiple skillsets (front end, back end, ops) that they aren't experts in. And plenty of bad code and technical debt from bad off/on-shore contractors due to skill issues and communication barriers.

I'm not saying you're wrong though, but just that it won't be much different.

It fascinates me how little respect software engineers get. The entire internet paints a picture of software engineers as useless code producing monkeys.

I haven’t seen anything similar in any other engineering industry.

Because in other industries the engineers know something about the real world thing that they're engineering, they aren't just manipulating abstractions.

Computers aren't real

I've seen people estimate two days of work for a manual headline content update on a mature code base they know.

No new tests, no new design, no new implementation. Literally just a single word swapped.

Turns out the engineerinng team had uncertainty with the new testing framework that was mandated, and rather than document it they blew up their velocity.

This approach bites back. When work isn't understood, people on the outside disrespect the workers.

Isn’t it a bit suspect that that particular image benefits a special new tool about which the OP is heavily speculating?

I am very surprised that no one in HN can see the correlation of the narrative and speculations vs the absurd age thrown around only to be repeated by actual people whose image is being tarnished because the mass narrative wants them to feel powerless(first stage of wage suppression).

Expert in debugging AI coding crap, I guess... People that still really understand code...

10 years is just too far out to talk intelligently about? I’d start with three years.

please don't worry, because AI, there will be new jobs created by AI, abundance will be in sight.

"Source: Trust Me, Bro"

Is this written with an LLM?

no clue

We're on the cusp of an indie software boom.

The reasons why exist in Taleb's antifragility thesis: the antifragile will gain from disorder.

The nature of the tech industry, in the decades roughly since the Cold War ended(which put to rest a certain pattern of tech focused on the MIC and moved SV forward into its leadership position), has promoted fragility along several of Taleb's dimensions: it aims to scale, it aims to centralize, and it aims to intervene. The pinnacle achievement of this trend is probably the iPhone, a convergent do-everything device that promises to change your life.

But it's axiomatic(in Taleb's view, which I will defer to since his arguments are good enough for me) that this won't last, and with talk of "the end of the US empire", and a broader pessimism in tech, there seems to be popular agreement that we are done with the scale narrative. AI is a last holdout for that narrative since it justifies further semiconductor investment, stokes national security fears of an "AI race" and so on - it appeals to forces with big pocketbooks, that are also big in scale, and also in a position of fragility themselves. But eventually they will tap out too, for the same reasons. Whether that's a "next year" thing or a "twenty years" thing is hard to predict - the fall of the USSR was similarly hard to predict.

The things that are antifragile within software are too abstract to intentionally develop within a codebase, and are more a measure of philosophy and how one models the world with data - CollapseOS is at the extreme end of this, where the viewpoint is bacterial - "all our computing infrastructure is doomed if it is not maintainable by solo operators" - but there are intermediate points along it where the kinds of software that are needed are in that realm of a plugin to a large app or an extension to an existing framework, and development intentionally aims not to break outside that scope. That thesis agrees with the "small niches" view of things.

Since we have succeeded in putting computers into every place we can think of, many of the things we need to do with them do already have a conventional "way of doing it," but can exist in a standard and don't need to be formalized behind a platform and business model - and LLM stuff does actually have some role in serving that by being an 80% translator of intent, while a technical specialist has work in adding finish and polish. And that side of things agrees with going deeper into the stack.

I believe one pressing issue that's faced with this transition is the inclination brought from the corporate dev environment to create a generalized admixture of higher and lower level coding - to do everything as a JS app with WASM bits that are "rewritten in Rust" - when what indie software needs is more Turbo Pascals, Hypercards, and Visual Basics, environments that are tuned to be complete within themselves and towards the kinds of apps being written, while being "compatible enough" to deploy to a variety of end-user systems.

10 years is a long time.

Enough time for LLMs to see marked improvement and possibly the hallucination issue to be significantly reduced or solved.

When then happens the remaining software engineers won’t be skilled. Like many people in the principle or staff position or managerial position… technical skills don’t matter. AI now handles technical skills and the people controlling the AI are people who can navigate the politics.

That's more unqualified handwaving than in the rest of the entire thread. Technical skills will always matter. Office politics will always be useless.

If ai handles all the technical skills what’s left?

Like you made a claim here with no logical reasoning as to why.

My reasoning is simple if AI handles all technical skills then no technical skills are required for the job anymore. Where’s your reasoning?

In addition to this “unqualified hand waving”. Look at your own baseless statement. At least I qualified my statement with “it’s possible”.

If LLMs solve the hallucination issue then it’s stuff like this that will sit at the top of the hierarchy. People who make grand claims with confidence and play politics to get to the top. You say my statement is handwavy but really that’s another way of saying “I don’t need to prove you wrong I’m just going to make a baseless claim about how your statement is utter crap but do it in a way as to subvert HN politeness rules, i can say it in a way that doesn’t raise any eyebrows and people who already agree with me will automatically take my side even though I didn’t present any additional new reasoning here”.

This is what I’m talking about. It’s people who are good at strategies like this who will occupy the top spots in the future should LLMs continue to improve. Case in point. If karma was voting for the next person in a leadership position you would win.

> If AI handles all the technical skills what’s left?

This is a view utterly out of touch with reality. AI handling all technical skills? And who defines where technical skills start and end?

Will AI development/innovation stop in the future? And if not, will the engineers working on AI not be applying technical skills? Will AI eliminate the technical skill of plumbing? Of doctors? What about systems analysis and architectural design, a technical skill. Will AI read the minds of people, anticipate their every tech need, and eliminate that too?

Maybe some have a weird understanding of the "technical". Its meaning is much broader that they think.

Again how is it out of touch with reality? We have memes about vibe coders who don’t understand shit. Why do those memes exist? Because that’s where the trend-lines of technology are pointing. Whether we get there is another story. But denying the trendline is irrational.

Oh I just read your second paragraph. I’m obviously only talking about coding. Blue collar jobs that involves a lot of manual skill is less of an issue. Robotics has an upward trendline but it’s nowhere near moving at the breakneck pace of AI.

For now, the hallucination problem is objectively only getting worse with reasoning models, and there is no solution in sight.

Technical skills will matter for the simple reason that someone will always need to supervise the output of the AI. A non-technical person can't hope to do so too well. Without this supervision, the AI's work product will have skeletons that will cause unexpected issues due to outliers. Reality is full of outliers - it's what pays half the salary.

Imagine asking an AI to design an airplane. The design passes all test flights and also software tests. Would you want to fly in it without an expert human having reviewed the ins and outs of the design?

How about a CT scan machine? Would you want to risk the 10x radiation due to a hypothetical implementation error that strikes in 1 out of 10K cases?

> For now, the hallucination problem is objectively only getting worse with reasoning models, and there is no solution in sight.

You pulled this statement out of your ass. Objectively? We have baseline quantitative tests that say the opposite? LLMS are doing better on tests and it’s been improving. Where did your “objective” statement come from? Anecdotes? Quotations?

> Technical skills will matter for the simple reason that someone will always need to supervise the output of the AI.

Humans will never be so stupid as to lose all technical skill. In the beginning there needs to be mild technical skill at most so the human can somewhat understand what the AI is doing. But as trust grows the human will understand less and less of it. The human is like a desperate micro manager attempting to cling to understanding but that inevitably eventually erodes completely.

For commentary on the hallucination problem, please see https://news.ycombinator.com/item?id=43942800 and the associated news link. I have observed it myself in practice with o3 compared to o1.