I don't think it has ever been the case that you could neglect soft skills. You will hear this over and over, in every area of every business: people become successful by adjusting their behaviour to what works for the business. Sometimes this is called being a slick politician, sometimes it is called avoiding getting bogged down in politics.
But it's never been the case that a dev could just focus on technical things and not spend any time figuring out the context they are working in, and behaving accordingly.
My first day of work, this is what my boss said to me: "Look at this trading floor. There's screens everywhere, everything is numbers. Deltas, gammas, vegas. Everything is calculated by computers. But don't forget, every business is a people business!"
You’re right but i have always preferred people who can do a little more. Nothing against the socially awkward and conflict avoidant nature in many of these friends, but people who push back and fight to communicate their views and passions often got our team better outcomes than someone who just turns up and does the work they’re asked to do.
As long as it is not opposite set of skills (talks a lot without knowledge to back it up so essentially using charisma to convince people to do the wrong thing most of the time) then yes, a lil bit of negotiation can save you a whole lot of work in the long run (XY problems being one example)
For sure, I’ve been tricked into hiring those people before too. It’s good that there’s still something hard in running an organization, the whole “what is value?” question feels like it’ll be one of the few things we have to maintain work for humans over the next little while.
It definitely happens at bloated organizations that aren’t really good at software development. I think it is especially more common in organizations where software is a cost center and business rules involve a specialized discipline that software developers wouldn’t typically have expertise in.
I've heard this "soft skills are the only skills that matter" thing throughout my entire career but these days this is indeed greatly amplified.
Our CEO started sharing screenshots of his xitter/linkedin feeds and most of them contain wisdoms like "Opus 4.5 is better than 90% of talent". There is also longer form business fan fiction. It usually follows this template: there is a hero - a humble manager/sales person, and a villain - a cocky, nerdy software developer. The villain claims that some task is impossible or takes months to complete. Then the humble hero (equipped with Opus 4.5) completes the task in 2 hours. The villain is then humiliated/fired and everyone lives happily ever after.
These posts definitely contribute to the declining morale among employees. Nobody goes "above and beyond" anymore - we just strictly doing the tickets and nothing more.
Once upon a time a clever software engineer realized that engineering talent is the fuel which the business relies on to support its revenue growth, and management is for facilitating this process, while the CEO’s purpose is to be blamed when it doesn’t work out. He wrote a small bash script which replaced corporate leadership with a “quote of the day” generator and everyone lived happily ever after.
- You can have a horrible CEO that doesn't value their employees and is trying to devalue labor.
- AI coding tools can be incredible exoskeletons in the hands of skilled engineers and enable them to get much more work done.
Perhaps the real "SaaS-killer" is innovation capital [1] realizing it can take advantage of the various forms of arbitrage and changing of the guards happening now, raise venture capital, and take on the old and slow management-driven businesses.
If you've ever had the itch to fire your boss, now's the time. It's a hard path, there are way more hats to wear, but the dry powder is out there waiting to be deployed.
The sort of people who have been able to neglect their soft skills are probably still going to be able to do so in the future. But those were always the extreme intelligence outliers, generational talent type figures.
For most software engineers, neglect of soft skills have always been a career tarpit that leads nowhere you want to end up. Being able to navigate social settings and to communicate well is a force multiplier. For most people, it really doesn't matter how good you are if nobody understands what you are saying and you can't convince other people to buy into your ideas. You far more often see moderately successful charlatans that are all talk than successful people with awful communication skills. Of course if you're able to walk the talk, that's when you can really go places.
I think the author would say that the developer who is without soft skills won't merely be prevented from gaining desirable work. They'll be unable to keep a job, period.
Seems a pretty sketchy assertion, but regardless whether these people burn out in career purgatory at a java 8 feature factory moving jira tickets around for all of eternity, or they move on to something else entirely, it's probably not what they had in mind.
Doubt we'll see that in the short term. Long term, possibly, especially if you add a financial crisis.
Truth is most larger software development organizations could have even before LLMs downsized significantly and not lost much productivity.
The X formerly known as Twitter did this and has been chugging along on a fraction on its original staff count. It's had some brand problems since its acquisition, but those are more due to Mr Musk's eccentricities and political ventures than the engineering team.
The reason this hasn't happened to any wider degree is quarterly capitalism and institutional inertia. Looks weird to the investors if the organization claims to be doing well but is also slashing its employee count by 90%. Even if you bring a new CEO in that has these ideas, the org chart will fight it with tooth and nail as managers would lose reportees and clout.
Consultancies in particular are incredibly inefficient by design since they make more money if they take more time and bring a larger headcount to the task: They don't sell productivity, but man hours. Hence horrors like SAFe.
> Today, I use Claude Code for almost all non-trivial programming tasks and have spent $500+ on it just last December.
Truthfully, I don't want to get advice from people who become addicted to AI, sorry. The money investment that person did, already leaves me with tons of questions.
> Truthfully, I don't want to get advice from people who become addicted to AI, sorry.
If you sleep on this, these people are going to take your job.
I've been writing serious systems code for 15 years. Systems that handled billions of dollars of transaction volume a day and whose hourly outages cost billions of dollars. These are systems you have to design carefully. Active-active, beyond five nines reliable.
I'm telling you AI is extremely beneficial even in this segment of the market. The value prop is undeniable.
I'm easily getting twice my workload done with AI, and I'm not even leveraging the full extent of the tools. I've only just started to do more than fancy tab-autocomplete.
This is going to be a huge shift in our industry, and I would brace for impact.
Be careful, engineers, when interacting with soft skill experts not to join their reality distortion field where it’s all about coordination, alignment, bizness strategy, clever planning. Whereas the real stuffs are just implémentation details, quickly solved.
The trouble isn't that they would convince me with their reality distortion (they don't), the issue is that they are satisfied with their "progress" while I'm still asking crucial questions (which they ignore, b/c they don't see the importance).
Part of the soft skills that are useful for a developer is being able to frame your crucial questions in a way that makes their importance relevant to their world.
Wait until you try to explain basic concepts to such people! What's annoying is HR employing more of these people and not understanding why the dial does not move.
What is baffling and infuriating is when such people are put in management. MBAs will destroy Western business eventually(those who tag an MBA onto a STEM qualification are not as bad but still infected.)
Software 'engineers' are going to have learn the hard skills that they have gotten by without thus far. The ones who were never engineers in the first place will have to learn 'soft skills' because they are incapable. I believe @ Tsoding put it best: The only thing AI will do, in so far as coding goes, is to remove a lot of people who should never have been in it in the first place.
I hope this too but it's not a given, IMO. Previously people without technical chops failed quickly by being unable to deliver working code, now they can deliver mediocre code with the damage only becoming clear years later. It breaks the "can deliver code --> good technical ability" proxy and even after the initial damage wave, it's unclear if we will find a better proxy.
My rule of thumb is that if you’re not measuring anything you’re not engineering. It’s not the whole picture, but to me the engineering part sometimes means being able to explain (and even quantify) why one solution is better than another.
I've found coding assistants to be a huge boon for this. All of the thorough analysis that previously would've taken a bunch of tedious extra thought work to do for marginal benefit (with a well-calibrated intuition) becomes 5 seconds of thought to the the computer to build a harness and then letting it chew on that for 15 minutes. It now also takes me one command and less than a minute to get pprof captures from all the production services my team owns (thanks to some scripts I had it write), which is just something I never would've bothered to automate otherwise, so we never really looked much at it. Codex is also very good at analyzing the results, and finding easy wins vs. knowing what would be invasive to improve, and then just doing it.
Thinking of seeing if I can get mutation testing set up next, and expanding our use of fuzzing. All of these techniques that I know about but haven't had the time to do are suddenly more feasible to invest into.
Not the original author, but I would guess that understanding the domain problem and interpreting it correctly in a software solution (not code, but a product with workflows, UX, etc.), which in turn requires ability to listen and understand and ask right questions on one hand (what a user wants to achieve), and a good understanding of the technical limitations as well as human habits on the other hand (what is possible and makes sense). One can argue that AI lacks what we'd call intuition and interpersonal qualities which are still necessary, as before AI.
Your $300k+ TC job is going away. The only way you'll make the same take home is if you provide more value.
You can be a robotic IC, but you won't be any better than a beginner with Claude Code. You have to level up your communication and organizational value to stay at the top.
Everyone has to wear the cloth of a senior engineer now to simply stay in place. If you can't write well, communicate well, plan and organize, you're not providing more value than a Claude-enhanced junior.
"If you can't write well, communicate well, plan and organize"
Straw man. Pretty sure, this is the dilbert equivalent of "I can problem solve". If you are an engineer, we are making boatloads being brought in to fix the incompetence of this level of thinking. INFOSEC alone is having a field day.
Would you like to buy a bridge? Coded by Claude. One previous owner. An owner who used said bridge to go to church once a week, and vibe code in Starbucks afterwoods.
Beyond engineering itself, strictly computer engineering?
How many coders have no idea what goes on behind an IDE. Have not even the slightest notion how a computer works. Who thinks building a computer means watching a Youtube video and buying ready made parts, putting them together, and then think they should be employed by NASA.
To begin: Math, Linux, Devops, C, and Assembly.
Not a youtube video. Not arithmetic. Learn to the point that you could be employed by any of the above as a senior. And don't fear failure. Keep doing it until you understand it.
I agree with your original post that the need for hard skills will persist, but I see it in the other direction: software engineers are going to have to get better at thinking in larger abstractions, not deeper understanding of the stack. Those who can only solve problems locally and repeat the patterns they've seen before rather than create new patterns from building blocks are the ones who are going to struggle.
"software engineers are going to have to get better at thinking in larger abstractions"
........Math was first on my list.
I don't know how else to say that.
Yep, exactly. The failure to realize that you mean different things when talking about "larger abstractions" is exactly the kind of miscommunication that software people will need to navigate better in the future.
Computer science is indistinguishable from sufficiently advanced maths.
The AI can already do that part.
The abstraction that matters going forward, is understanding why the abstraction chosen by the AI does or doesn't match the one needed by the customer's "big picture".
The AI is a bit too self-congratulatory in that regard, even if it can sometimes spot its own mistakes.
A lot of studying math is just learning jargon and applications for what are actually pretty straightforward concepts, which lets you better communicate with the computer. You get higher bandwidth communication and better ability to know all of the nuances in things it might propose. You can propose things and understand when it replies with nuances you missed.
Like intro differential geometry is basically a deep dive into what one actually does when reading a paper map. Something everyone (over 30?) is familiar with. But it turns out there's plenty to fill a graduate level tome on that topic.
Linear algebra is basically studying easy problems: y=ax. Plenty to write about how to make your problem (or at least parts of it) fit that mould.
I suspect and think I've seen others say that you get better outputs from LLMs when using jargon. Essentialy, its pattern matching tells it to say what an expert would say when using the terminology experts use.
Ah, I think “Math” as a single word on its means many different things to many different people, I didn’t interpret in quite the same way. But I see what you mean.
I’m not sure that my colleagues who I think of as “good at math” and “good at thinking in larger abstractions” are necessarily the same ones, but there’s definitely a lot of overlap.
I think this is kinda sad. I always thought coding was so beautiful for creating a (good paying) job for socially awkward people on the spectrum who pre-coding was considered dorks and went into idk what but probably some unsatisfying job. It seems like we're heading back to that which makes me sad and worried about the highly functioning autistic population.
I am sorry but is being employed a job or a daycare?
Not only this is extremely patronizing towards all people on spectrum, but at the same time extremely hurtful statement for people who are treating employment as a job(ie - most of population).
And what are you going to say to people who are stuck in low-end jobs?
I thought this article was going to be about something else ...
It is really about prompting and writing specs - the "soft" (but really "hard") skill of giving detailed specs to an LLM so it does what you want.
I think the more important, truly soft, skill in the age of AI is going to be communicating with humans and demonstrating your value in communicating both vertically up and down and horizontally within your team. LLMs are becoming quite capable at the "autistic" skill of coding, but they are still horrible communicators, and don't communicate at all unless spoken to. This is where humans are currently, and maybe for a long time, irreplaceable - using our soft skills to interact with other humans and as always translate and refine fuzzy business requirements into the unforgiving language of the machine, whether than is carefully engineered LLM contexts, or machine code.
As far as communication goes, I have to say that Gemini 3.0, great as it is, is starting to grate on me with it's sycophantic style and failure to just respond as requested rather than to blabber on about "next steps" that it is constantly trying to second guess from it's history. You can tell it to focus and just answer the question, but that only lasts for one or two conversational turns.
One of Gemini's most annoying traits is to cheerfully and authoritatively give design advice, then when questioned admit (or rather tell, as if it were it's own insight) that this advice is catastrophically bad and will lead to a bad outcome, and without pause then tell you what you really should do, as if this is going to be any better.
"You're absolutely right! You've just realized the inevitable hard truth that all designers come to! If you do [what I just told you to do], program performance will be terrible! Here is how you avoid that ... (gives more advice pulled out of ass, without any analysis of consequences)"
So I opened this article to find out at the very beginning, that author put's a lot of money to AI providers... So probably also used it to write this article. So, according to the rule "text which is not worth of spending time on writing is also no worth of reading" I closed it.
I found out that the author of a blog post paid a maid to clean their house, and sends their laundry out. Therefore, the blog post was written by the maid, or one of the laundromat employees. So I closed it.
That's a false equivalence. You might read a blog post about engineering from someone who paid a maid to clean their house, and sends their laundry out.
But would you read their blog post on laundry tips?
No - it's just as easy for you to send out your own laundry.
Err, no, the equivalent would be if the author had written an article about laundry, and how important it is to pay attention to all the details... then you realize the photo of his laundry still has a tag on it from the laundromat.
I'm just going to sharpen up my hard skills so that I don't have to suck up on my soft skills. If that doesn't work out as I wish, well, since I already have a job, and I'm already 40+, and it looks like my company doesn't go away in a few years and we are so short-handed, I think I'm fine.
Just to clarify that I'm not a jackass in real-life. In fact, I'm perfectly OK with all sorts of soft skills -- after all, my current position requires me to do so. But I just try to maintain a minimum level of soft skills to navigate the shoreline -- not interested to move up anyway.
"I'm already 40+, and it looks like my company doesn't go away in a few years and we are so short-handed"
FANG employee for over a decade quit her job, opened a consultancy, and sold back what she was doing to former clients of her company(and their clients). She sold back what she was having to do to make up for this short-handedness, and the incompetence of many of her former colleagues.
She did it at 3-400 times the markup she was being paid while employed :) because they were time critical.
I mentor CS students at two local universities. The best students are using gen ai to enhance their learning and understanding (i.e. they use it as a tool instead of a crutch). The worst students are using it in attempt to “level the playing field” and are failing miserably.
It is easy to determine if someone solved a problem using AI because they can’t explain or recreate “their” solution. Detecting cheating in essays is still far more difficult.
You could probably detect essay cheating (AI written) in the exact same way by questioning the student about it - why did they organize the essay in this way, what was their motivation for focusing on X, or expressing something as Y... Of course anyone can concoct an explanation on the fly, but it should be obvious if they are speaking from the experience of having authored it or just coming up with a post-hoc rationalization.
If they had AI write the essay, yet can still explain it as well as if they had written it themselves (ditto for code), then it would tend to indicate that they at least read it and thought about it, which I think should be more acceptable in a learning environment.
This couldn't ring more true to me - I think one of the consequences of the rapid change in the profession we are seeing is that skills that typically were required only at more senior levels become required further down the stack.
If I was a junior today, I'd be studying business impact, effective communication, project management, skills that were previously something you could get away with under-indexing on until senior+.
What about those professional software developers still refusing to use AI / LLMs? I know a couple and they're still churning out code completely 100% manually.
Heck, I even know a guy who refuses to use an IDE with Java and the indenting is a mess, but he gets there.
Not necessarily faster, but more easy for sure. There's plenty of stories of proficient abacus using accountants being faster than those using calculators. Those days are gone now though because a calculator is just so much easier to pick up.
In a team environment, half of the job is communication.
That LLMs do a better job if you know what you are asking for is old news.
But to be honest, I usually don't care to write properly into an LLM prompt. An LLM will ignore grammar and form and just extract the essence. If I make an actual mistake I will notice quickly and fix it.
If I'd send slack messages like that to an peer, they'd either mock me or simply think I am dumb.
We also know the stories about people that use LLMs for any communication or anything they write. Probably for the exact reason that being lazy with writing is acceptable now. My call is that writing skill will decline, not improve. This could probably be the case for anything that people use LLMs as a proxy for.
Broadly agree, but one point I think is (sadly) relevant:
> That LLMs do a better job if you know what you are asking for is old news.
Even a decade after Word Lens had demonstrated augmented reality live translation through a smartphone camera, I was amazing people by showing them the same feature in Google Translate.
Similar anecdotes about Shakuntala Devi, even in 2018 I was seeing claims about her mental arithmetic beating a supercomputer (claims that ignored that this happened in 1977 and the computer was already obsolete at the time), even though my mid-2013 MacBook Air could not only beat her by a factor of 150 million, it could also train an AI to read handwritten numbers from scratch in 0.225 seconds, and then perform inference (read numbers) at just over 6,629 digits per second*.
You say "old news", I say this discussion will be on repeat even in the early 2030s. And possibly even the 2060s.
I think AI coding agents will quickly pick all the low hanging fruits and plateau.
Everything can be vibed will be vibed until everyone hits a wall, where no docs to form corpus nor instructions for prompts exist. There are problems that are yet to be named, but how can you name things when humans aren't the one to experience patterns of a thought process?
And naming things is one of the only two hard things in computer science
Counter point - If one person can now do the work of an entire team the level of communication skills required will actually be simplified.
So now instead of needing to manage multiple stakeholders and expectations of 10 different middle managers you'll probably just have a 1:1 with a single person.
I would submit that this could be based on a stereotype that “coder = antisocial.”
Why are we assuming that people who write code don’t have soft skills?
The youngest generation who joined the profession are probably in it for the the salary versus the older generations who came from computer clubs and dungeons and dragons groups of the 1970s/1980s along with a culture where having a niche interest was socially ostracizing and uncool.
I wonder if the youngest generation entering the profession is much more of a cross section of regular people?
Look, I personally am taking full advantage of exactly the skills described. I was the one who posted the above thing on HN showing how I am 20-50x more productive now, complete with a 4 hour speedrun video. I usually try not to just talk and point out current problems, but build solutions AND show (github, youtube) with specific details so you can watch it and apply it for yourself. But I am telling you:
1) most people will not adapt, so we will need UBI for those who don’t
2) eventually even those who adapt will be replaced too, so we will need UBI for everybody
It is after all a thin layer that remains. I remember Kasparov proudly talked about how “centaurs” (human + machine working together) in chess were better than machines alone… until they weren’t, and human in the loop became a liability.
But the problem is more widespread in the last 70 years. Just look around. Industry always tells the individual they can do some individual action downstream to clean up the mess they create upstream, and it is leading the entire planet into ruin:
In fact, the human population in modern environments has been living large on an ecological credit card and the bill is coming due for our children, because all the “individual responsibility” stuff — where you can somehow diet, exercise and recycle your way out of things corporations do upstream — is all a gient lie and always has been. So the negative externalities just build up until the next generation won’t be able to ignore them anymore, but it could be too late. Whether that’als day zero for water in cities, or factory farms for meat with antiobiotic resistance, or fossil fuels and greenhouse gases to subsidize the car industry, or ubiqitous microplastic plastic pollution around thr world (yes, personal plastic recycling was just another such scam designed to keep you docile and not organize to force corporations to switch to biodegradeable materials.) The “anthoposcene” is seeing a decline in insects and all species of animal except humans and farm animals. Coral reefs are bleached, kelp forests and rainforests are decimated, and governments work with industry to eg allow Patagonian forests to be burned for new developments and then smokey the bear says “only YOU can prevent forest fires”. Think about it.
I may have misread your comment, but I don't think soft skills are a 'narrow thing' at all. Effective communication, building trust, bringing people along with you - these are fundamental to being an effective human, not some niche pivot.
Not fair on you. I did not mean to have a dig. I get where you are coming from, and should have elaborated. I've worked with those one or two engineers who were rude by default. Who had an extraordinary knack of vaguely describing the problem set, and then having a full on meltdown, always in front of other people, when the solutions did not match the problem in their head.*
*Goldman Sachs(sorry for invoking that name here) did a report on their high turnover, and the above framing was why many quit.
Look, if we zoom in, then "learning to code" is also quite a broad range of skills that someone needs to master before they can meaningfully carve out a career in a competitive marketplace.
The point is that if you zoom out, it's just a thin slice that can be automated by machines. People keep saying "I'll tell you in my experience, no UAV will ever trump a pilot's instinct, his insight, the ability to look into a situation beyond the obvious and discern the outcome, or a pilot's judgment"... https://www.youtube.com/watch?v=ZygApeuBZdk
But as you can see, they're all wrong. By narrow here I meant a thin layer that thinks it's indispensable as they remove all the other layers. Until the system comes for this layer too.
I don't think it has ever been the case that you could neglect soft skills. You will hear this over and over, in every area of every business: people become successful by adjusting their behaviour to what works for the business. Sometimes this is called being a slick politician, sometimes it is called avoiding getting bogged down in politics.
But it's never been the case that a dev could just focus on technical things and not spend any time figuring out the context they are working in, and behaving accordingly.
My first day of work, this is what my boss said to me: "Look at this trading floor. There's screens everywhere, everything is numbers. Deltas, gammas, vegas. Everything is calculated by computers. But don't forget, every business is a people business!"
There are plenty of devs who do nothing beyond taking a Jira ticket scoped by others, implementing it, and then grabbing the next ticket.
While they may not have been very successful, they did have a place.
You’re right but i have always preferred people who can do a little more. Nothing against the socially awkward and conflict avoidant nature in many of these friends, but people who push back and fight to communicate their views and passions often got our team better outcomes than someone who just turns up and does the work they’re asked to do.
As long as it is not opposite set of skills (talks a lot without knowledge to back it up so essentially using charisma to convince people to do the wrong thing most of the time) then yes, a lil bit of negotiation can save you a whole lot of work in the long run (XY problems being one example)
For sure, I’ve been tricked into hiring those people before too. It’s good that there’s still something hard in running an organization, the whole “what is value?” question feels like it’ll be one of the few things we have to maintain work for humans over the next little while.
Is this genuinely common? I’ve only ever seen that level of hand holding extended to new grad hires.
It definitely happens at bloated organizations that aren’t really good at software development. I think it is especially more common in organizations where software is a cost center and business rules involve a specialized discipline that software developers wouldn’t typically have expertise in.
Often it means being a sociopath
I've heard this "soft skills are the only skills that matter" thing throughout my entire career but these days this is indeed greatly amplified.
Our CEO started sharing screenshots of his xitter/linkedin feeds and most of them contain wisdoms like "Opus 4.5 is better than 90% of talent". There is also longer form business fan fiction. It usually follows this template: there is a hero - a humble manager/sales person, and a villain - a cocky, nerdy software developer. The villain claims that some task is impossible or takes months to complete. Then the humble hero (equipped with Opus 4.5) completes the task in 2 hours. The villain is then humiliated/fired and everyone lives happily ever after.
These posts definitely contribute to the declining morale among employees. Nobody goes "above and beyond" anymore - we just strictly doing the tickets and nothing more.
Once upon a time a clever software engineer realized that engineering talent is the fuel which the business relies on to support its revenue growth, and management is for facilitating this process, while the CEO’s purpose is to be blamed when it doesn’t work out. He wrote a small bash script which replaced corporate leadership with a “quote of the day” generator and everyone lived happily ever after.
Two things can be true:
- You can have a horrible CEO that doesn't value their employees and is trying to devalue labor.
- AI coding tools can be incredible exoskeletons in the hands of skilled engineers and enable them to get much more work done.
Perhaps the real "SaaS-killer" is innovation capital [1] realizing it can take advantage of the various forms of arbitrage and changing of the guards happening now, raise venture capital, and take on the old and slow management-driven businesses.
If you've ever had the itch to fire your boss, now's the time. It's a hard path, there are way more hats to wear, but the dry powder is out there waiting to be deployed.
[1] ICs in both senses of the acronym.
The sort of people who have been able to neglect their soft skills are probably still going to be able to do so in the future. But those were always the extreme intelligence outliers, generational talent type figures.
For most software engineers, neglect of soft skills have always been a career tarpit that leads nowhere you want to end up. Being able to navigate social settings and to communicate well is a force multiplier. For most people, it really doesn't matter how good you are if nobody understands what you are saying and you can't convince other people to buy into your ideas. You far more often see moderately successful charlatans that are all talk than successful people with awful communication skills. Of course if you're able to walk the talk, that's when you can really go places.
I think the author would say that the developer who is without soft skills won't merely be prevented from gaining desirable work. They'll be unable to keep a job, period.
Seems a pretty sketchy assertion, but regardless whether these people burn out in career purgatory at a java 8 feature factory moving jira tickets around for all of eternity, or they move on to something else entirely, it's probably not what they had in mind.
The shift is from tarpit to unemployment. A Jira ticket processing dev still has use. Probably not for much longer.
Doubt we'll see that in the short term. Long term, possibly, especially if you add a financial crisis.
Truth is most larger software development organizations could have even before LLMs downsized significantly and not lost much productivity.
The X formerly known as Twitter did this and has been chugging along on a fraction on its original staff count. It's had some brand problems since its acquisition, but those are more due to Mr Musk's eccentricities and political ventures than the engineering team.
The reason this hasn't happened to any wider degree is quarterly capitalism and institutional inertia. Looks weird to the investors if the organization claims to be doing well but is also slashing its employee count by 90%. Even if you bring a new CEO in that has these ideas, the org chart will fight it with tooth and nail as managers would lose reportees and clout.
Consultancies in particular are incredibly inefficient by design since they make more money if they take more time and bring a larger headcount to the task: They don't sell productivity, but man hours. Hence horrors like SAFe.
> Today, I use Claude Code for almost all non-trivial programming tasks and have spent $500+ on it just last December.
Truthfully, I don't want to get advice from people who become addicted to AI, sorry. The money investment that person did, already leaves me with tons of questions.
> Truthfully, I don't want to get advice from people who become addicted to AI, sorry.
If you sleep on this, these people are going to take your job.
I've been writing serious systems code for 15 years. Systems that handled billions of dollars of transaction volume a day and whose hourly outages cost billions of dollars. These are systems you have to design carefully. Active-active, beyond five nines reliable.
I'm telling you AI is extremely beneficial even in this segment of the market. The value prop is undeniable.
I'm easily getting twice my workload done with AI, and I'm not even leveraging the full extent of the tools. I've only just started to do more than fancy tab-autocomplete.
This is going to be a huge shift in our industry, and I would brace for impact.
Be careful, engineers, when interacting with soft skill experts not to join their reality distortion field where it’s all about coordination, alignment, bizness strategy, clever planning. Whereas the real stuffs are just implémentation details, quickly solved.
In the end, they might convince you that 2+2=5.
The trouble isn't that they would convince me with their reality distortion (they don't), the issue is that they are satisfied with their "progress" while I'm still asking crucial questions (which they ignore, b/c they don't see the importance).
Part of the soft skills that are useful for a developer is being able to frame your crucial questions in a way that makes their importance relevant to their world.
Wait until you try to explain basic concepts to such people! What's annoying is HR employing more of these people and not understanding why the dial does not move. What is baffling and infuriating is when such people are put in management. MBAs will destroy Western business eventually(those who tag an MBA onto a STEM qualification are not as bad but still infected.)
Software 'engineers' are going to have learn the hard skills that they have gotten by without thus far. The ones who were never engineers in the first place will have to learn 'soft skills' because they are incapable. I believe @ Tsoding put it best: The only thing AI will do, in so far as coding goes, is to remove a lot of people who should never have been in it in the first place.
I hope this too but it's not a given, IMO. Previously people without technical chops failed quickly by being unable to deliver working code, now they can deliver mediocre code with the damage only becoming clear years later. It breaks the "can deliver code --> good technical ability" proxy and even after the initial damage wave, it's unclear if we will find a better proxy.
Can you elaborate on this?
What are some examples of skills you think are now essential, that prior have been taken for granted or obviated in some way?
My rule of thumb is that if you’re not measuring anything you’re not engineering. It’s not the whole picture, but to me the engineering part sometimes means being able to explain (and even quantify) why one solution is better than another.
I've found coding assistants to be a huge boon for this. All of the thorough analysis that previously would've taken a bunch of tedious extra thought work to do for marginal benefit (with a well-calibrated intuition) becomes 5 seconds of thought to the the computer to build a harness and then letting it chew on that for 15 minutes. It now also takes me one command and less than a minute to get pprof captures from all the production services my team owns (thanks to some scripts I had it write), which is just something I never would've bothered to automate otherwise, so we never really looked much at it. Codex is also very good at analyzing the results, and finding easy wins vs. knowing what would be invasive to improve, and then just doing it.
Thinking of seeing if I can get mutation testing set up next, and expanding our use of fuzzing. All of these techniques that I know about but haven't had the time to do are suddenly more feasible to invest into.
Not the original author, but I would guess that understanding the domain problem and interpreting it correctly in a software solution (not code, but a product with workflows, UX, etc.), which in turn requires ability to listen and understand and ask right questions on one hand (what a user wants to achieve), and a good understanding of the technical limitations as well as human habits on the other hand (what is possible and makes sense). One can argue that AI lacks what we'd call intuition and interpersonal qualities which are still necessary, as before AI.
Read further into the comment.
Your $300k+ TC job is going away. The only way you'll make the same take home is if you provide more value.
You can be a robotic IC, but you won't be any better than a beginner with Claude Code. You have to level up your communication and organizational value to stay at the top.
Everyone has to wear the cloth of a senior engineer now to simply stay in place. If you can't write well, communicate well, plan and organize, you're not providing more value than a Claude-enhanced junior.
"If you can't write well, communicate well, plan and organize" Straw man. Pretty sure, this is the dilbert equivalent of "I can problem solve". If you are an engineer, we are making boatloads being brought in to fix the incompetence of this level of thinking. INFOSEC alone is having a field day.
Would you like to buy a bridge? Coded by Claude. One previous owner. An owner who used said bridge to go to church once a week, and vibe code in Starbucks afterwoods.
> If you can't write well, communicate well, plan and organize,
Why not ask the LLM to write for you? Same for planning, organization and written communication.
Seems like robotic ICs can "robotize" most of the work stack.
Beyond engineering itself, strictly computer engineering? How many coders have no idea what goes on behind an IDE. Have not even the slightest notion how a computer works. Who thinks building a computer means watching a Youtube video and buying ready made parts, putting them together, and then think they should be employed by NASA.
To begin: Math, Linux, Devops, C, and Assembly. Not a youtube video. Not arithmetic. Learn to the point that you could be employed by any of the above as a senior. And don't fear failure. Keep doing it until you understand it.
I agree with your original post that the need for hard skills will persist, but I see it in the other direction: software engineers are going to have to get better at thinking in larger abstractions, not deeper understanding of the stack. Those who can only solve problems locally and repeat the patterns they've seen before rather than create new patterns from building blocks are the ones who are going to struggle.
"software engineers are going to have to get better at thinking in larger abstractions" ........Math was first on my list. I don't know how else to say that.
> I don't know how else to say that.
Yep, exactly. The failure to realize that you mean different things when talking about "larger abstractions" is exactly the kind of miscommunication that software people will need to navigate better in the future.
If you need to have that explained to you, you are the problem.
Computer science is indistinguishable from sufficiently advanced maths.
The AI can already do that part.
The abstraction that matters going forward, is understanding why the abstraction chosen by the AI does or doesn't match the one needed by the customer's "big picture".
The AI is a bit too self-congratulatory in that regard, even if it can sometimes spot its own mistakes.
A lot of studying math is just learning jargon and applications for what are actually pretty straightforward concepts, which lets you better communicate with the computer. You get higher bandwidth communication and better ability to know all of the nuances in things it might propose. You can propose things and understand when it replies with nuances you missed.
Like intro differential geometry is basically a deep dive into what one actually does when reading a paper map. Something everyone (over 30?) is familiar with. But it turns out there's plenty to fill a graduate level tome on that topic.
Linear algebra is basically studying easy problems: y=ax. Plenty to write about how to make your problem (or at least parts of it) fit that mould.
I suspect and think I've seen others say that you get better outputs from LLMs when using jargon. Essentialy, its pattern matching tells it to say what an expert would say when using the terminology experts use.
Ah, I think “Math” as a single word on its means many different things to many different people, I didn’t interpret in quite the same way. But I see what you mean.
I’m not sure that my colleagues who I think of as “good at math” and “good at thinking in larger abstractions” are necessarily the same ones, but there’s definitely a lot of overlap.
I think this is kinda sad. I always thought coding was so beautiful for creating a (good paying) job for socially awkward people on the spectrum who pre-coding was considered dorks and went into idk what but probably some unsatisfying job. It seems like we're heading back to that which makes me sad and worried about the highly functioning autistic population.
On the other hand, the increase in remote roles has made this a bit easier for some.
I am sorry but is being employed a job or a daycare?
Not only this is extremely patronizing towards all people on spectrum, but at the same time extremely hurtful statement for people who are treating employment as a job(ie - most of population).
And what are you going to say to people who are stuck in low-end jobs?
I'm on the spectrum to be clear
I thought this article was going to be about something else ...
It is really about prompting and writing specs - the "soft" (but really "hard") skill of giving detailed specs to an LLM so it does what you want.
I think the more important, truly soft, skill in the age of AI is going to be communicating with humans and demonstrating your value in communicating both vertically up and down and horizontally within your team. LLMs are becoming quite capable at the "autistic" skill of coding, but they are still horrible communicators, and don't communicate at all unless spoken to. This is where humans are currently, and maybe for a long time, irreplaceable - using our soft skills to interact with other humans and as always translate and refine fuzzy business requirements into the unforgiving language of the machine, whether than is carefully engineered LLM contexts, or machine code.
As far as communication goes, I have to say that Gemini 3.0, great as it is, is starting to grate on me with it's sycophantic style and failure to just respond as requested rather than to blabber on about "next steps" that it is constantly trying to second guess from it's history. You can tell it to focus and just answer the question, but that only lasts for one or two conversational turns.
One of Gemini's most annoying traits is to cheerfully and authoritatively give design advice, then when questioned admit (or rather tell, as if it were it's own insight) that this advice is catastrophically bad and will lead to a bad outcome, and without pause then tell you what you really should do, as if this is going to be any better.
"You're absolutely right! You've just realized the inevitable hard truth that all designers come to! If you do [what I just told you to do], program performance will be terrible! Here is how you avoid that ... (gives more advice pulled out of ass, without any analysis of consequences)"
It's getting old.
So I opened this article to find out at the very beginning, that author put's a lot of money to AI providers... So probably also used it to write this article. So, according to the rule "text which is not worth of spending time on writing is also no worth of reading" I closed it.
I found out that the author of a blog post paid a maid to clean their house, and sends their laundry out. Therefore, the blog post was written by the maid, or one of the laundromat employees. So I closed it.
That's a false equivalence. You might read a blog post about engineering from someone who paid a maid to clean their house, and sends their laundry out.
But would you read their blog post on laundry tips?
No - it's just as easy for you to send out your own laundry.
Err, no, the equivalent would be if the author had written an article about laundry, and how important it is to pay attention to all the details... then you realize the photo of his laundry still has a tag on it from the laundromat.
false equivalence
I'm just going to sharpen up my hard skills so that I don't have to suck up on my soft skills. If that doesn't work out as I wish, well, since I already have a job, and I'm already 40+, and it looks like my company doesn't go away in a few years and we are so short-handed, I think I'm fine.
Just to clarify that I'm not a jackass in real-life. In fact, I'm perfectly OK with all sorts of soft skills -- after all, my current position requires me to do so. But I just try to maintain a minimum level of soft skills to navigate the shoreline -- not interested to move up anyway.
"I'm already 40+, and it looks like my company doesn't go away in a few years and we are so short-handed" FANG employee for over a decade quit her job, opened a consultancy, and sold back what she was doing to former clients of her company(and their clients). She sold back what she was having to do to make up for this short-handedness, and the incompetence of many of her former colleagues.
She did it at 3-400 times the markup she was being paid while employed :) because they were time critical.
I’m not in FAANG and mine is just a small one which may not afford a lot, but yeah I totally agree it’s a good idea for a FAANG person.
The irony here is that universities are struggling to teach writing skills, due to massive cheating with AI.
I mentor CS students at two local universities. The best students are using gen ai to enhance their learning and understanding (i.e. they use it as a tool instead of a crutch). The worst students are using it in attempt to “level the playing field” and are failing miserably.
It is easy to determine if someone solved a problem using AI because they can’t explain or recreate “their” solution. Detecting cheating in essays is still far more difficult.
You could probably detect essay cheating (AI written) in the exact same way by questioning the student about it - why did they organize the essay in this way, what was their motivation for focusing on X, or expressing something as Y... Of course anyone can concoct an explanation on the fly, but it should be obvious if they are speaking from the experience of having authored it or just coming up with a post-hoc rationalization.
If they had AI write the essay, yet can still explain it as well as if they had written it themselves (ditto for code), then it would tend to indicate that they at least read it and thought about it, which I think should be more acceptable in a learning environment.
This couldn't ring more true to me - I think one of the consequences of the rapid change in the profession we are seeing is that skills that typically were required only at more senior levels become required further down the stack.
If I was a junior today, I'd be studying business impact, effective communication, project management, skills that were previously something you could get away with under-indexing on until senior+.
What about those professional software developers still refusing to use AI / LLMs? I know a couple and they're still churning out code completely 100% manually.
Heck, I even know a guy who refuses to use an IDE with Java and the indenting is a mess, but he gets there.
The classic "technically strong but lacking soft skills" unheard story about you through the office after your results.
Counter-point, a three person dev team displaces a multi million revenue company.
Why have a slow human CEO when machines are faster..
In education, I view them as life skills, durable skills.
calculators doing a perfect job did not end accountants' jobs, it made it faster.
I don’t think that’s as good of an example as you think it is.
“Computer” used to be a job title. It was entirely replaced by … drumroll … electronic computers, i.e. calculators.
Technology doesn’t usually eliminate the need for a job output in general but it can sometimes shift the skills needed wildly.
Not necessarily faster, but more easy for sure. There's plenty of stories of proficient abacus using accountants being faster than those using calculators. Those days are gone now though because a calculator is just so much easier to pick up.
In a team environment, half of the job is communication.
That LLMs do a better job if you know what you are asking for is old news.
But to be honest, I usually don't care to write properly into an LLM prompt. An LLM will ignore grammar and form and just extract the essence. If I make an actual mistake I will notice quickly and fix it. If I'd send slack messages like that to an peer, they'd either mock me or simply think I am dumb. We also know the stories about people that use LLMs for any communication or anything they write. Probably for the exact reason that being lazy with writing is acceptable now. My call is that writing skill will decline, not improve. This could probably be the case for anything that people use LLMs as a proxy for.
Broadly agree, but one point I think is (sadly) relevant:
> That LLMs do a better job if you know what you are asking for is old news.
Even a decade after Word Lens had demonstrated augmented reality live translation through a smartphone camera, I was amazing people by showing them the same feature in Google Translate.
Similar anecdotes about Shakuntala Devi, even in 2018 I was seeing claims about her mental arithmetic beating a supercomputer (claims that ignored that this happened in 1977 and the computer was already obsolete at the time), even though my mid-2013 MacBook Air could not only beat her by a factor of 150 million, it could also train an AI to read handwritten numbers from scratch in 0.225 seconds, and then perform inference (read numbers) at just over 6,629 digits per second*.
You say "old news", I say this discussion will be on repeat even in the early 2030s. And possibly even the 2060s.
* Uses an old version of python, you'll need to fix it up accordingly: https://benwheatley.github.io/blog/2018/03/16-10.44.18.html
I think AI coding agents will quickly pick all the low hanging fruits and plateau.
Everything can be vibed will be vibed until everyone hits a wall, where no docs to form corpus nor instructions for prompts exist. There are problems that are yet to be named, but how can you name things when humans aren't the one to experience patterns of a thought process?
And naming things is one of the only two hard things in computer science
You never really could. If you hear of a (very) succesful software engineer with horrid soft skills they’re a 1%er chance wise
I’ve been hearing this since the 90’s.
I feel like it's a myth promoted by engineers with poor soft skills to justify themselves never learning said skills.
Counter point - If one person can now do the work of an entire team the level of communication skills required will actually be simplified.
So now instead of needing to manage multiple stakeholders and expectations of 10 different middle managers you'll probably just have a 1:1 with a single person.
I would submit that this could be based on a stereotype that “coder = antisocial.”
Why are we assuming that people who write code don’t have soft skills?
The youngest generation who joined the profession are probably in it for the the salary versus the older generations who came from computer clubs and dungeons and dragons groups of the 1970s/1980s along with a culture where having a niche interest was socially ostracizing and uncool.
I wonder if the youngest generation entering the profession is much more of a cross section of regular people?
As usual - the advice is essenially rats from a sinking ship all the way. “You all need to do this narrow thing to survive now”.
2016 to truckers: “Learn to code LOL”
2026 to coders: “Learn soft skills”
https://news.ycombinator.com/item?id=46436872
Look, I personally am taking full advantage of exactly the skills described. I was the one who posted the above thing on HN showing how I am 20-50x more productive now, complete with a 4 hour speedrun video. I usually try not to just talk and point out current problems, but build solutions AND show (github, youtube) with specific details so you can watch it and apply it for yourself. But I am telling you:
1) most people will not adapt, so we will need UBI for those who don’t
2) eventually even those who adapt will be replaced too, so we will need UBI for everybody
It is after all a thin layer that remains. I remember Kasparov proudly talked about how “centaurs” (human + machine working together) in chess were better than machines alone… until they weren’t, and human in the loop became a liability.
But the problem is more widespread in the last 70 years. Just look around. Industry always tells the individual they can do some individual action downstream to clean up the mess they create upstream, and it is leading the entire planet into ruin:
https://magarshak.com/blog/government-and-industry-distract-...
In fact, the human population in modern environments has been living large on an ecological credit card and the bill is coming due for our children, because all the “individual responsibility” stuff — where you can somehow diet, exercise and recycle your way out of things corporations do upstream — is all a gient lie and always has been. So the negative externalities just build up until the next generation won’t be able to ignore them anymore, but it could be too late. Whether that’als day zero for water in cities, or factory farms for meat with antiobiotic resistance, or fossil fuels and greenhouse gases to subsidize the car industry, or ubiqitous microplastic plastic pollution around thr world (yes, personal plastic recycling was just another such scam designed to keep you docile and not organize to force corporations to switch to biodegradeable materials.) The “anthoposcene” is seeing a decline in insects and all species of animal except humans and farm animals. Coral reefs are bleached, kelp forests and rainforests are decimated, and governments work with industry to eg allow Patagonian forests to be burned for new developments and then smokey the bear says “only YOU can prevent forest fires”. Think about it.
I may have misread your comment, but I don't think soft skills are a 'narrow thing' at all. Effective communication, building trust, bringing people along with you - these are fundamental to being an effective human, not some niche pivot.
"Effective communication, building trust, bringing people along with you" That's a David Brent powerpoint presentation.
Fair. I'll retire 'bringing people along with you' before it ends up on a motivational poster with a stock photo of a rowing team.
Though you're right that there's no I in team. There is one in AI though, which probably tells us something.
Not fair on you. I did not mean to have a dig. I get where you are coming from, and should have elaborated. I've worked with those one or two engineers who were rude by default. Who had an extraordinary knack of vaguely describing the problem set, and then having a full on meltdown, always in front of other people, when the solutions did not match the problem in their head.*
*Goldman Sachs(sorry for invoking that name here) did a report on their high turnover, and the above framing was why many quit.
Look, if we zoom in, then "learning to code" is also quite a broad range of skills that someone needs to master before they can meaningfully carve out a career in a competitive marketplace.
The point is that if you zoom out, it's just a thin slice that can be automated by machines. People keep saying "I'll tell you in my experience, no UAV will ever trump a pilot's instinct, his insight, the ability to look into a situation beyond the obvious and discern the outcome, or a pilot's judgment"... https://www.youtube.com/watch?v=ZygApeuBZdk
But as you can see, they're all wrong. By narrow here I meant a thin layer that thinks it's indispensable as they remove all the other layers. Until the system comes for this layer too.