I really love using AI to code but more and more I wonder ... Are things really that different? So I guess I'm the business as usual type.
I think on the frontend side we're going to see a lot more scope for teams.
On a backend infra side it seems as hard as ever. Still have to think really hard problems, think deeply about data structure and flow, and deal with second- and third-order effects. Or even harder because the models like to confidently lie.
The harder question is how we train people but that doesn't seem insurmountable either. Most of us cut our teeth as junior engineers somewhere, implementing tasks that Claude can now do without breaking a sweat but was that really the most efficient way to train and learn?
I work on backend infra. I touch a lot of things and have multiple instances of claude code running at once. I do feel like I'm doing more simultaneously. but that's still more that I have to keep track of and make sense of in my head. claude is making me way more productive, for sure, but at the same time I feel way more overwhelmed than I've ever been.
idk how anyone else is doing it and managing all of this. supposedly there's people with large teams of agents that they can just trust to do everything end-to-end.
I’ve been finding I can mitigate the sense of being overwhelmed by trying to keep my tasks within the same domain or part of the code base. I know it’s not always an option, though.
I guess the goal is to stack context rather than spread it.
I find it incredibly exhausting otherwise. It’s a skill I don’t have in spades. I’ll have to develop it, though.
I don’t think the people trusting agents end to end are getting the quality they think they are. I’m bullish on AI as a permanent and genuinely useful tool in our kit, but I’m not seeing signs that you can actually let them loose at all. Looping on a very well defined problem, sure. But open-ended tasks across large infrastructures and complex domains, no, it’s going to be duct tape and sprawl as far as the IDE can see. It’s going to sprawl faster than context limits can grow, and the mess will only get worse.
Right there with you - this has been my experience as well, to a tee
I can do more now…so I do - it’s really that simple
And it’s way more exhausting because there’s no room to breathe - fresh code to work with every few minutes with prepping for the next set of tasks in between
On the one hand the dopamine’s got me hooked on this line a video
On the other…I’m as overwhelmed as ever line you said, even if it’s my own doing
My feelings right now can be described as follows:
If AI progress stalls now or grinds to a halt, we get to keep a lot of new jobs that are going to show up (it opens a lot of doors), while maintaining software devs as an interesting career.
Egoistically, I would love that.
If it keeps going and software devs get replaced, so many jobs are going to disappear.
Even though the synthetic benchmarks paint a picture of LLMs coming a long way since 2022, my practical experience has been that they aren’t tangibly better. No doubt someone reading this will chime in and say LLMs are way better at writing code or whatever, and maybe that’s true, but there’s no difference between ChatGPT 3.5 and Claude Opus 4.8 as far as my trusting the output. Opus 4.8 still messes up plenty. It’s particularly bad with identifying and fixing CI yaml, but it struggles in the usual areas too.
So I’m thinking we’ve just about reached apex with LLMs, and they have failed at replacing software engineers (companies can freeze hiring juniors at their own, future peril using any excuse they like).
In my experience I'd have to agree. I'm shipping more and I'm onboarding onto domains faster but the same bottlenecks exist, the same complexities creep up and the same talents that help individuals push through are still relevant.
What I like is it's still hard, but you can do things like "prove this using game theory" or "find the optimal value for this and a proof" without having to know game theory or deep math really
LLMs seem to be best at writing web apps intended to be internal tools. I don’t need to ever really read the code because the functionality is relatively simple- but still valuable. I had Claude Code build me a CLI tool for running common kubectl commands against our EKS cluster. With EKS (AWS) you generally use the AWS CLI to authenticate and then choose the correct EKS cluster and then you can run your kubectl commands. So this CLI tool remembers all the AWS accounts and EKS cluster names and namespaces and pods that I care about. It’s been a huge time saver, and I never would have had the time to build it as part of my day to day. And because it’s an internal tool, I don’t have to worry about things like security/authentication, and any bugs aren’t as big a deal as if they were customer-facing. I wouldn’t want to use it for generating an app that is going on the public internet and doing anything sensitive/important/valuable with my or other people’s data (I’m sure that doesn’t stop others from vibe coding commercial products).
Yeah I mean, as far as I can tell the result of the agent mania is the same amount of software, but an acceleration in the decline of performance and quality. I'm also increasingly seeing early adopters going back to a more "traditional" approach to development. So idk, maybe the result will be more jobs fixing up all the vibe code while we transition to a more mature implementation of language models into our workflows.
Reversing to which direction? Because what I've always seen here is a pretty good mix of positive and negative sentiments. Usually we get a lot of AI related submissions, but with skeptics/opposers in the comments.
I’m not sure. I’ve been reading death-of-the-software engineer for years, but recently the -vibe- feels different. I don’t have anything anecdotal to back it up so take it with a grain of salt. I might be reading what I want to see
I'm assuming it's a turn to the negative and not more positivity you're seeing? Geohot's article and Hasimoto's tweet about AI psychosis kind of made me pay attention.
I've been both types on and off and on for my whole career. At times, very engrossed in the technology itself - but I got started because I needed tools for my job that didn't exist so I learned how to build them. I've also developed and maintained open source software - some of it very much for its own sake as a technology - some of it quite utilitarian.
Understanding of product and business has always differentiated me though, and this is why I've never really stressed about any of this.
Another thing I've noticed - most developers are really bad at reviewing code - whether AI wrote it or not. Its really hard to make your brain sink in deep enough to really evaluate what you are looking at. I think a lot of developers never - or almost never - find bugs based on code inspection alone. Once they are written - there is often no other practical way to confirm that tests actually test what they claim to other than inspection. And bugs in design are still very costly in this new world.
As long as any human still has an edge in any aspect of the software development process - people who can force themselves to really think through proposed designs, test plans and cases and code will be really valuable.
But has software engineering really changed that much?
I've barely had any regular fulltime work experience most of my career has been contracting, so I’m not sure.It seems like the perceived change varies depending on the environmen
What I don’t get is using AI agents feels basically the same as what I used to do with legacy codebases of 100k–200k lines—understanding the codebase and making partial fixes or adding features. To me, it feels like nothing has changed.
Most of my career has been about finding parts that won't break within the overall code structure—without necessarily knowing the entire detailed specification—and adding features or fixing bugs there. So I feel like it's no different from finding and fixing small issues and bugs in a large codebase written by AI.
Of course, when delivering a solution that's 70,000–80,000 lines long, there is a change things that used to rely on templates and CMS tools can now be created more diversely using AI. But aside from that, I don't think things have changed as much as people say. It might be different for those who build things entirely from scratch with AI, though.
My code writing ability has declined, but I'm not really seeing a dramatic change in workflow
Now I work with code written by AI, adding features and modifying it…
Most of the codebases I’ve seen were bad anyway—there was no good code to begin with. When coding with AI, I split tasks into P0, P1, P2 based on importance. For P0, I write everything myself. For P1, I write the draft and AI implements it. For P2, AI implements everything. For P0, I only handle things that involve responsibility, like payment logic or login logic.
I don't participate in open source, so I don’t see big changes. The only thing I notice is that AI speeds things up a lot—for things I used to understand by reading documentation and examples, now I can generate them much faster. Personally, I wish open source projects had more simple, short examples in their code
I really liked reading through the Mars trilogy. It imagined a world where AI is used for fluent effortless translation - local languages get a renaissance since now you _dont_ need a lengua Franca, everyone just speaks what they like the most, and can understand everyone else. Much more “flavour” to human interaction.
Also ai makes things just resource constrained, not labour - whatever you imagined, you could make happen, just needed to “talk to an ai” about it. Lots of terraforming Mars / Venus in that book were imagined like that.
But it also analysed the social / political / behavioural aspects of it. Places that had to preserve old power structures - aka US/Europe/China - got engulfed with mega corps controlling everything etc.
But Mars - where people had enough freedom to imagine something different, came up with political/financial structures to incorporate all of that, and thrived.
I think it tried to play the card of “if US was being created right now - what would its ideals be” If you had a huge tract of land that was “free” and nobody (powerful enough) claiming it, and a population that didn’t yet have strong allegiances and could be persuaded to band together, what would AI, tech and all these years of progress allow us as humans to achieve politically.
Which also makes you feel kinda sad for the US in that world - it is the old rusted power center that can’t innovate and is stuck in the past…
Now it’s only sci fi of course, but it was quite interesting to imagine a world where AI gets smarter and smarter but never reaches that “sentient” threshold. I think the whole trilogy aged incredibly well all things considered.
The issues are really social/political/behavioural.
If white collar jobs are decimated the economy will collapse, because it's still - nominally - a consumer economy, even though most of the spending comes from the top 10%.
If 25% of that top 10% is suddenly out of work, that's a huge chunk of GDP gone, and a likely depression as the knock on effects spread.
I don't see anyone thinking about this seriously. [Something something UBI] is not a serious attempt to deal with the problem.
Ultimately you end up with an economy where most humans are surplus to economic requirements and will literally be murdered by economics.
Or wealth and opportunity are widely distributed, but "jobs" are nothing like they are today.
In the old days we programmed systems by literally wiring them. There wasn't much work, only a few "programmers" were employed. Then somebody came up with Punch cards that was much more vision than wiring the systems directly. This opened the door for a lot more people to use them and now programmers were busier.
The punch cards didn't scale either so eventually we created panels with buttons so we could type the programs into the computers. That was more efficient and now all the sudden it lowered the bar entry and more people who are employed and doing the work.
Assembly language to machine code compilers to assembly language high-level languages and LLMs.
Every time developing software gets easier, it only increases the amount of work required. I'm busier today than I've ever been in my entire life.
What's going to happen to software engineers? They're going to be overworked and they're going to be given more work and the cycle will never end.
Once upon a time there were engineers that used software. Just like any other tool, and usually in combination with electronic, electrical and mechanical equipment, all of them being very well aware of the laws governing it all.
But it was so great as a tool that some engineers didn't want to deal with the burdens and limitations of the physical world, and started focusing on software more and more.
Then the software engineers came, for whom the physical and mathematical aspect of the whole thing was just a distant history lesson (and preferably a problem in someone else's computer).
And after software engineers, the only constant in the entire ordeal will remain: engineering, in a shape or form that very likely nobody can predict right now.
As someone not old but young enough to not have experienced a world before software, Im not sure that engineering rooted in the physical is a necessary prerequisite.
I appreciate your comment but the entire world I happened to experience in my coming of age was at the dawn of the consumer internet. And so “web stuff” was how I cut my teeth. And its my profession. And i never went to school for it, im basically a dumb untrained web dev, borne from the script kiddie days.
There’s a stigma to it sure, but im well past it. All to say I just dont think CS principles down to the physics level is the root and all is an abstraction. Theyre just different things now.
I doubt we will ever need a large team to build software again. PMs will be fractional. Product owners and subject matter experts will become more valuable. And engineers that have deep experience with things like Domain-Driven Design will thrive and vibe coders will eventually be shut out of bigger roles.
Traditionally those roles of providing "alignment" to overcome organizational inertia were held by program managers. So that seems like IC roles are indeed going away, to be replaced by more "technical program manager" type roles.
Maybe it's my experience, but TPMs were often responsible for coordinating large org-wide or cross-org initiatives. It's prohibitively expensive to have TPMs on anything smaller.
Since engineering with AI is still very technical, I would wager that software engineers would stretch into less technical areas of software development rather than TPMs stretching into technical areas. I only say this as someone with experience with AI and I see how easy it is to write bad code with AI if you're not aware of what it's doing.
OP here. So I've also been running a survey about how developers are feeling about AI (added a link to the post) and 60% of respondents said that using AI actively makes them more tired. Most also report being more productive with AI. One would think these two things combined make up for initiatives for shorter work weeks and the like, but from what I've seen, at least for now, the opposite seems to be happening.
P.S. Sample size is currently small but I've also gathered a fair amount of qualitative data to indicate both of these things. Of course measuring productivity is also far from a science.
Shorter work weeks don't mean less work though. Not from the point of view of management, at least. It just means that 5 days are now compressed to 4, probably making the problem much worse and burn out much more likely.
In my experience, this is what ended up happening.
I've definitely noticed a distinct lack of pride now that Claude Code is writing 90% of the code I'm delivering these days. For simple problems (which most are) it works well enough and you are definitely shipping code faster - and with actual test coverage to boot. But it just doesn't feel the same - there's little craftsmanship and honestly it's boring as fuck. You spend a lot of time setting up guardrails and having it produce plans that you then have to refactor multiple times. It's impressive that LLMs can do this, but it's not particularly enjoyable. I guess I was a "writing code was the fun part" guy.
Semi-related, but I really want to see the long term maintenance outcomes of all code being produced by these software engineers that were apparently just closeted project managers. I feel like having 50% of the engineers in this industry just telling Claude Code, "yeah that looks good to me" 150 times a day is going to result in an incredible amount of software rewriting.
I wonder that too. I've been on the receiving end of "it's 90% done, we just need someone to get it over the line for us" way too many times to know that there's going to be a lot of pain trying to maintain or re-write parts of anything that is vibe-coded.
On the other hand, I notice the AI-fundamentalists(I am not sure how to refer to people within that group) just say that you won't be doing any hand coding anymore and you'd "just" ask something like claude to maintain it or re-write.
AI made working for (an AI-pilled) company unbearable to me. Time will tell if there will be new approaches that new companies take, and lessons learned.
But immediately, I hate AI work in a company setting. Mandates intentionally want humans managing more “features”. When you get stuck owning someone else’s AI “feature”, it’s game over.
The new hope is that as a founder, the stimulating and creative parts of working with AI are persevered. Fingers crossed so far so good.
My feelings about front-end code are that I have a stronger feeling of craftsmanship, in the sense that I can ship a much more polished product because all the small nagging annoyances are things that I can eliminate (second-hand), where I’d previously have just lived with a lot of them as fixing them took too long to be worthwhile. I hate shipping some of the resulting working slop.
For me, part of craftsmanship is the quality of the shipped product. (I’m also willing to use CNC tools while doing hobby woodworking; others think that takes away the craftsmanship; I think it changes how the craftsmanship is experienced and applied.)
There are people who make handmade watches in Switzerland and people who push buttons and trigger automations on assembly lines for mass manufactured watch companies in Vietnam.
The guy in Switzerland enjoys his life more, their labor has more leverage and is generally more valued across the world.
You all have a choice. Handmade software isnt going anywhere.
In fact this is all a trick to decrease the bargaining power of your labor, turn you into a de-skilled button pusher and a slave to anyone willing to pay you a morsel. Dont fall for it.
Edit: The people arguing with this have low standards and yearn for slop.
I think there's a ton of examples where this is true for lower level stuff like open source where you see the internals.
For commerical products it certainly exists too, for example in those cases where you know the product is built by one person or a small group of people who you absolutely know take extraordinary care to get all the details right, and it shows through as a really nice intangible feeling when you're using the product.
That (kind of rare to be honest) 'oh this is just really well done' feeling.
Platforms that are obviously vibe slop are almost entirely ignored by those that value quality.
If your product has that cheap stench of AI I'll never trust it, I'll instantly think the creators are chumps and I'll look for any alternative that doesn't have that stench.
Fortunately the economics of infinitely copyable software aren't going anywhere. If there is actually a sizable market, it can be served with relatively little cost, compared to hand-producing physical goods. I am not fully convinced there will be such a market, though.
Its like telling animals dont get domesticated, dont get exploited, dont get eaten, think about the farmer, be more like him.
Its not going to happen. For the first time in history we have tools that talk back. Adults arent mentally prepped for that. They will cling to old models of reality they know, when reality has changed. The kids on the other hand can learn and adapt better, but learning takes time and happens through trial and error.
All anyone in a position of responsibility can do, just like parents, protect the kids as much as you can as they keep falling and trust with time they will learn to adapt to the new reality.
The difference, and I think we as an industry will have to reconcile this depending on how advanced llms get, is that you don’t see the quality in handmade code like you do in a high end watch or a luxury automobile or appliance. The veneer might be identical. It’s going to be tough to convince people that handmade software has added value or quality over slop. I still believe it does right now, but that might not always be true. And this is an industry that has pumped out a lot of sloppy code for decades, even before it was actual slop.
well this is not exactly unique to software. It is not a given that the 'handmade' nature of a product (luxury or otherwise) manifests in anything tangible or self-evidently superior. Luxury products in general overwhelmingly treat the crafted nature of the product as almost solely an investment in narrative (read: marketing) and market positioning, not an actual material outcome.
What're you talking about? I can instantly detect if software was created with AI. It has a stench much like a cheap product who's materials off gassing, and Im certainly not using it over the handmade solution. There will always be two options, like in clothes or watches.
Nothing. There will be more of us than ever as the amount of money each engineer can generate goes up. I’m not sure why everyone comes to the conclusion that there will be less engineers as productivity increases. We aren’t that expensive relative to the value we generate and AI, when wielded by an expert, will exponentially increase the amount of return a single engineer can generate.
LLMs have taken much of the enjoyment out of coding for me, but I don’t think it has to be that way. I hope that as an industry we settle on LLMs being more like tools than human assistants. I think most of the engineers at my company are using LLMs as human assistants- most of them have agentic workflows set up and have premium subscriptions to Claude AND Gemini AND ChatGPT. Many have local LLMs running on their company MacBook Pros, but they can never manage to describe to me in plain English what those local LLMs are doing. I would compare local LLMs to Raspberry Pi clusters. They’re neat, and technically they can do stuff but they are incredibly impractical. I want so desperately to have the power of Claude Opus running locally, but we are incredibly, ridiculously, extremely far off (benchmarks are often misleading and nobody likes to talk about the paltry tokens/s they’re getting on their home rig).
So the problem for me is that I’ve noticed a trend with my coworkers- they usually don’t have a formal CS background (e.g., Comp Sci degree), their resumes are unimpressive, and they have personal websites where anyone can download a copy of their resume. I haven’t seen this at any previous job. It’s just weird, man. Anyway, so my problem is when they are purported to be the best of the best at this company, and they’re chugging the LLM kool-aid, I’m struggling to figure out whether they’re really ahead of the curve and getting the drop on the rest of us, or they’re doing basically no work and waving through each other’s PRs with little to no oversight on what the LLMs are writing. The bush league nature of many of the bugs I’m seeing going all the way to prod makes me think it’s the latter.
Meanwhile, these guys and their teams are completing an absurd number of story points each sprint. My team is dead last (and coincidentally we’re all pretty anti-AI for writing code).
I’m waiting for the other shoe to drop. I’m waiting for the C suite to call our project a failure and lay everyone off. The product generally works, though I think if the devs were truly rockstars, things would be much smoother. It seems like LLMs are letting middle-of-the-road or mediocre devs appear to be rockstars purely on output speed. Nobody in this org cares if the output has 4x the bugs it should have, just that it was completed quickly.
The main issue I think is that nobody- including the rockstar tech leads- knows how the system works. They don’t know the code. There’s a bug and it takes forever to track down. Don’t ask about unit tests.
I think in a perfect world, these LLMs are treated like another tool rather than another person- meaning you and I still write a large portion of the code, and we read and understand every line- especially the lines generated by LLMs. It’s still a huge productivity multiplier for me to not have to remember how to do mundane things like write out arrays into CSV files- before I’d find a close implementation on StackOverflow and tweak to my needs. Now I can just describe what I want the CSV to look like and Claude Code generally gets it right on the first try.
I really love using AI to code but more and more I wonder ... Are things really that different? So I guess I'm the business as usual type.
I think on the frontend side we're going to see a lot more scope for teams.
On a backend infra side it seems as hard as ever. Still have to think really hard problems, think deeply about data structure and flow, and deal with second- and third-order effects. Or even harder because the models like to confidently lie.
The harder question is how we train people but that doesn't seem insurmountable either. Most of us cut our teeth as junior engineers somewhere, implementing tasks that Claude can now do without breaking a sweat but was that really the most efficient way to train and learn?
I work on backend infra. I touch a lot of things and have multiple instances of claude code running at once. I do feel like I'm doing more simultaneously. but that's still more that I have to keep track of and make sense of in my head. claude is making me way more productive, for sure, but at the same time I feel way more overwhelmed than I've ever been.
idk how anyone else is doing it and managing all of this. supposedly there's people with large teams of agents that they can just trust to do everything end-to-end.
I’ve been finding I can mitigate the sense of being overwhelmed by trying to keep my tasks within the same domain or part of the code base. I know it’s not always an option, though.
I guess the goal is to stack context rather than spread it.
I find it incredibly exhausting otherwise. It’s a skill I don’t have in spades. I’ll have to develop it, though.
I don’t think the people trusting agents end to end are getting the quality they think they are. I’m bullish on AI as a permanent and genuinely useful tool in our kit, but I’m not seeing signs that you can actually let them loose at all. Looping on a very well defined problem, sure. But open-ended tasks across large infrastructures and complex domains, no, it’s going to be duct tape and sprawl as far as the IDE can see. It’s going to sprawl faster than context limits can grow, and the mess will only get worse.
Right there with you - this has been my experience as well, to a tee
I can do more now…so I do - it’s really that simple
And it’s way more exhausting because there’s no room to breathe - fresh code to work with every few minutes with prepping for the next set of tasks in between
On the one hand the dopamine’s got me hooked on this line a video
On the other…I’m as overwhelmed as ever line you said, even if it’s my own doing
My feelings right now can be described as follows:
If AI progress stalls now or grinds to a halt, we get to keep a lot of new jobs that are going to show up (it opens a lot of doors), while maintaining software devs as an interesting career.
Egoistically, I would love that.
If it keeps going and software devs get replaced, so many jobs are going to disappear.
Even though the synthetic benchmarks paint a picture of LLMs coming a long way since 2022, my practical experience has been that they aren’t tangibly better. No doubt someone reading this will chime in and say LLMs are way better at writing code or whatever, and maybe that’s true, but there’s no difference between ChatGPT 3.5 and Claude Opus 4.8 as far as my trusting the output. Opus 4.8 still messes up plenty. It’s particularly bad with identifying and fixing CI yaml, but it struggles in the usual areas too.
So I’m thinking we’ve just about reached apex with LLMs, and they have failed at replacing software engineers (companies can freeze hiring juniors at their own, future peril using any excuse they like).
In my experience I'd have to agree. I'm shipping more and I'm onboarding onto domains faster but the same bottlenecks exist, the same complexities creep up and the same talents that help individuals push through are still relevant.
What I like is it's still hard, but you can do things like "prove this using game theory" or "find the optimal value for this and a proof" without having to know game theory or deep math really
LLMs seem to be best at writing web apps intended to be internal tools. I don’t need to ever really read the code because the functionality is relatively simple- but still valuable. I had Claude Code build me a CLI tool for running common kubectl commands against our EKS cluster. With EKS (AWS) you generally use the AWS CLI to authenticate and then choose the correct EKS cluster and then you can run your kubectl commands. So this CLI tool remembers all the AWS accounts and EKS cluster names and namespaces and pods that I care about. It’s been a huge time saver, and I never would have had the time to build it as part of my day to day. And because it’s an internal tool, I don’t have to worry about things like security/authentication, and any bugs aren’t as big a deal as if they were customer-facing. I wouldn’t want to use it for generating an app that is going on the public internet and doing anything sensitive/important/valuable with my or other people’s data (I’m sure that doesn’t stop others from vibe coding commercial products).
Yeah I mean, as far as I can tell the result of the agent mania is the same amount of software, but an acceleration in the decline of performance and quality. I'm also increasingly seeing early adopters going back to a more "traditional" approach to development. So idk, maybe the result will be more jobs fixing up all the vibe code while we transition to a more mature implementation of language models into our workflows.
I’ve noticed a massive reverse in AI sentiment in the last 3 weeks here on HN.
It’s not that I don’t disagree, but I wonder what’s going on. Maybe it’s the IPO
Reversing to which direction? Because what I've always seen here is a pretty good mix of positive and negative sentiments. Usually we get a lot of AI related submissions, but with skeptics/opposers in the comments.
I’m not sure. I’ve been reading death-of-the-software engineer for years, but recently the -vibe- feels different. I don’t have anything anecdotal to back it up so take it with a grain of salt. I might be reading what I want to see
I'm assuming it's a turn to the negative and not more positivity you're seeing? Geohot's article and Hasimoto's tweet about AI psychosis kind of made me pay attention.
https://news.ycombinator.com/item?id=48263238
https://xcancel.com/mitchellh/status/2055380239711457578
I've been both types on and off and on for my whole career. At times, very engrossed in the technology itself - but I got started because I needed tools for my job that didn't exist so I learned how to build them. I've also developed and maintained open source software - some of it very much for its own sake as a technology - some of it quite utilitarian.
Understanding of product and business has always differentiated me though, and this is why I've never really stressed about any of this.
Another thing I've noticed - most developers are really bad at reviewing code - whether AI wrote it or not. Its really hard to make your brain sink in deep enough to really evaluate what you are looking at. I think a lot of developers never - or almost never - find bugs based on code inspection alone. Once they are written - there is often no other practical way to confirm that tests actually test what they claim to other than inspection. And bugs in design are still very costly in this new world.
As long as any human still has an edge in any aspect of the software development process - people who can force themselves to really think through proposed designs, test plans and cases and code will be really valuable.
But has software engineering really changed that much? I've barely had any regular fulltime work experience most of my career has been contracting, so I’m not sure.It seems like the perceived change varies depending on the environmen
What I don’t get is using AI agents feels basically the same as what I used to do with legacy codebases of 100k–200k lines—understanding the codebase and making partial fixes or adding features. To me, it feels like nothing has changed.
Most of my career has been about finding parts that won't break within the overall code structure—without necessarily knowing the entire detailed specification—and adding features or fixing bugs there. So I feel like it's no different from finding and fixing small issues and bugs in a large codebase written by AI.
Of course, when delivering a solution that's 70,000–80,000 lines long, there is a change things that used to rely on templates and CMS tools can now be created more diversely using AI. But aside from that, I don't think things have changed as much as people say. It might be different for those who build things entirely from scratch with AI, though.
My code writing ability has declined, but I'm not really seeing a dramatic change in workflow
Now I work with code written by AI, adding features and modifying it… Most of the codebases I’ve seen were bad anyway—there was no good code to begin with. When coding with AI, I split tasks into P0, P1, P2 based on importance. For P0, I write everything myself. For P1, I write the draft and AI implements it. For P2, AI implements everything. For P0, I only handle things that involve responsibility, like payment logic or login logic.
I don't participate in open source, so I don’t see big changes. The only thing I notice is that AI speeds things up a lot—for things I used to understand by reading documentation and examples, now I can generate them much faster. Personally, I wish open source projects had more simple, short examples in their code
I really liked reading through the Mars trilogy. It imagined a world where AI is used for fluent effortless translation - local languages get a renaissance since now you _dont_ need a lengua Franca, everyone just speaks what they like the most, and can understand everyone else. Much more “flavour” to human interaction.
Also ai makes things just resource constrained, not labour - whatever you imagined, you could make happen, just needed to “talk to an ai” about it. Lots of terraforming Mars / Venus in that book were imagined like that.
But it also analysed the social / political / behavioural aspects of it. Places that had to preserve old power structures - aka US/Europe/China - got engulfed with mega corps controlling everything etc.
But Mars - where people had enough freedom to imagine something different, came up with political/financial structures to incorporate all of that, and thrived.
I think it tried to play the card of “if US was being created right now - what would its ideals be” If you had a huge tract of land that was “free” and nobody (powerful enough) claiming it, and a population that didn’t yet have strong allegiances and could be persuaded to band together, what would AI, tech and all these years of progress allow us as humans to achieve politically.
Which also makes you feel kinda sad for the US in that world - it is the old rusted power center that can’t innovate and is stuck in the past…
Now it’s only sci fi of course, but it was quite interesting to imagine a world where AI gets smarter and smarter but never reaches that “sentient” threshold. I think the whole trilogy aged incredibly well all things considered.
The issues are really social/political/behavioural.
If white collar jobs are decimated the economy will collapse, because it's still - nominally - a consumer economy, even though most of the spending comes from the top 10%.
If 25% of that top 10% is suddenly out of work, that's a huge chunk of GDP gone, and a likely depression as the knock on effects spread.
I don't see anyone thinking about this seriously. [Something something UBI] is not a serious attempt to deal with the problem.
Ultimately you end up with an economy where most humans are surplus to economic requirements and will literally be murdered by economics.
Or wealth and opportunity are widely distributed, but "jobs" are nothing like they are today.
Given where we are now option 1 looks closer.
They're going to be overworked.
In the old days we programmed systems by literally wiring them. There wasn't much work, only a few "programmers" were employed. Then somebody came up with Punch cards that was much more vision than wiring the systems directly. This opened the door for a lot more people to use them and now programmers were busier.
The punch cards didn't scale either so eventually we created panels with buttons so we could type the programs into the computers. That was more efficient and now all the sudden it lowered the bar entry and more people who are employed and doing the work.
Assembly language to machine code compilers to assembly language high-level languages and LLMs.
Every time developing software gets easier, it only increases the amount of work required. I'm busier today than I've ever been in my entire life.
What's going to happen to software engineers? They're going to be overworked and they're going to be given more work and the cycle will never end.
Once upon a time there were engineers that used software. Just like any other tool, and usually in combination with electronic, electrical and mechanical equipment, all of them being very well aware of the laws governing it all.
But it was so great as a tool that some engineers didn't want to deal with the burdens and limitations of the physical world, and started focusing on software more and more.
Then the software engineers came, for whom the physical and mathematical aspect of the whole thing was just a distant history lesson (and preferably a problem in someone else's computer).
And after software engineers, the only constant in the entire ordeal will remain: engineering, in a shape or form that very likely nobody can predict right now.
As someone not old but young enough to not have experienced a world before software, Im not sure that engineering rooted in the physical is a necessary prerequisite.
I appreciate your comment but the entire world I happened to experience in my coming of age was at the dawn of the consumer internet. And so “web stuff” was how I cut my teeth. And its my profession. And i never went to school for it, im basically a dumb untrained web dev, borne from the script kiddie days.
There’s a stigma to it sure, but im well past it. All to say I just dont think CS principles down to the physics level is the root and all is an abstraction. Theyre just different things now.
That's why I drink Genesee Light, to get drunk not for the flavor. I find the IPA drinkers to be of that same ilk
That is to say, somewhere along the way software got really complex, and really artistic, and really full of hubris.
I've been thinking about this since ChatGPT arrived.
My thoughts are distilled in a single page:
https://devarch.ai/trio-paradigm.html
I doubt we will ever need a large team to build software again. PMs will be fractional. Product owners and subject matter experts will become more valuable. And engineers that have deep experience with things like Domain-Driven Design will thrive and vibe coders will eventually be shut out of bigger roles.
> I doubt we will ever need a large team to build software again.
but that seems to contradict what you say here:
> vibe coders will eventually be shut out of bigger roles.
Isn't vibe coding driving that greater productivity and negating the need for large teams to build software?
Engineers are not going anywhere, they are going to fill the spaces outside of coding that are still critical to shipping new product.
Here is a post that summarizes what I mean: https://substack.com/home/post/p-200064883
Traditionally those roles of providing "alignment" to overcome organizational inertia were held by program managers. So that seems like IC roles are indeed going away, to be replaced by more "technical program manager" type roles.
Maybe it's my experience, but TPMs were often responsible for coordinating large org-wide or cross-org initiatives. It's prohibitively expensive to have TPMs on anything smaller.
Since engineering with AI is still very technical, I would wager that software engineers would stretch into less technical areas of software development rather than TPMs stretching into technical areas. I only say this as someone with experience with AI and I see how easy it is to write bad code with AI if you're not aware of what it's doing.
Four day working weeks would be nice.
OP here. So I've also been running a survey about how developers are feeling about AI (added a link to the post) and 60% of respondents said that using AI actively makes them more tired. Most also report being more productive with AI. One would think these two things combined make up for initiatives for shorter work weeks and the like, but from what I've seen, at least for now, the opposite seems to be happening.
P.S. Sample size is currently small but I've also gathered a fair amount of qualitative data to indicate both of these things. Of course measuring productivity is also far from a science.
Shorter work weeks don't mean less work though. Not from the point of view of management, at least. It just means that 5 days are now compressed to 4, probably making the problem much worse and burn out much more likely.
In my experience, this is what ended up happening.
I've definitely noticed a distinct lack of pride now that Claude Code is writing 90% of the code I'm delivering these days. For simple problems (which most are) it works well enough and you are definitely shipping code faster - and with actual test coverage to boot. But it just doesn't feel the same - there's little craftsmanship and honestly it's boring as fuck. You spend a lot of time setting up guardrails and having it produce plans that you then have to refactor multiple times. It's impressive that LLMs can do this, but it's not particularly enjoyable. I guess I was a "writing code was the fun part" guy.
Semi-related, but I really want to see the long term maintenance outcomes of all code being produced by these software engineers that were apparently just closeted project managers. I feel like having 50% of the engineers in this industry just telling Claude Code, "yeah that looks good to me" 150 times a day is going to result in an incredible amount of software rewriting.
I wonder that too. I've been on the receiving end of "it's 90% done, we just need someone to get it over the line for us" way too many times to know that there's going to be a lot of pain trying to maintain or re-write parts of anything that is vibe-coded.
On the other hand, I notice the AI-fundamentalists(I am not sure how to refer to people within that group) just say that you won't be doing any hand coding anymore and you'd "just" ask something like claude to maintain it or re-write.
AI made working for (an AI-pilled) company unbearable to me. Time will tell if there will be new approaches that new companies take, and lessons learned.
But immediately, I hate AI work in a company setting. Mandates intentionally want humans managing more “features”. When you get stuck owning someone else’s AI “feature”, it’s game over.
The new hope is that as a founder, the stimulating and creative parts of working with AI are persevered. Fingers crossed so far so good.
My feelings about front-end code are that I have a stronger feeling of craftsmanship, in the sense that I can ship a much more polished product because all the small nagging annoyances are things that I can eliminate (second-hand), where I’d previously have just lived with a lot of them as fixing them took too long to be worthwhile. I hate shipping some of the resulting working slop.
For me, part of craftsmanship is the quality of the shipped product. (I’m also willing to use CNC tools while doing hobby woodworking; others think that takes away the craftsmanship; I think it changes how the craftsmanship is experienced and applied.)
There are people who make handmade watches in Switzerland and people who push buttons and trigger automations on assembly lines for mass manufactured watch companies in Vietnam.
The guy in Switzerland enjoys his life more, their labor has more leverage and is generally more valued across the world.
You all have a choice. Handmade software isnt going anywhere.
In fact this is all a trick to decrease the bargaining power of your labor, turn you into a de-skilled button pusher and a slave to anyone willing to pay you a morsel. Dont fall for it.
Edit: The people arguing with this have low standards and yearn for slop.
People take pride in wearing handmade watches
As of today, I've never heard of anyone taking pride in using a SaaS or frequenting a website or an app because it was handcrafted. Maybe some day
I think there's a ton of examples where this is true for lower level stuff like open source where you see the internals.
For commerical products it certainly exists too, for example in those cases where you know the product is built by one person or a small group of people who you absolutely know take extraordinary care to get all the details right, and it shows through as a really nice intangible feeling when you're using the product.
That (kind of rare to be honest) 'oh this is just really well done' feeling.
You're completely wrong.
Platforms that are obviously vibe slop are almost entirely ignored by those that value quality.
If your product has that cheap stench of AI I'll never trust it, I'll instantly think the creators are chumps and I'll look for any alternative that doesn't have that stench.
They all have that stench.
I’m glad mass manufactured watches exist so the common man can afford them.
Fortunately the economics of infinitely copyable software aren't going anywhere. If there is actually a sizable market, it can be served with relatively little cost, compared to hand-producing physical goods. I am not fully convinced there will be such a market, though.
The Swiss govt has done much to protect Swiss labor from external market pressures; whether the watch is a handmade movement or quartz is irrelevant.
Its like telling animals dont get domesticated, dont get exploited, dont get eaten, think about the farmer, be more like him.
Its not going to happen. For the first time in history we have tools that talk back. Adults arent mentally prepped for that. They will cling to old models of reality they know, when reality has changed. The kids on the other hand can learn and adapt better, but learning takes time and happens through trial and error.
All anyone in a position of responsibility can do, just like parents, protect the kids as much as you can as they keep falling and trust with time they will learn to adapt to the new reality.
The difference, and I think we as an industry will have to reconcile this depending on how advanced llms get, is that you don’t see the quality in handmade code like you do in a high end watch or a luxury automobile or appliance. The veneer might be identical. It’s going to be tough to convince people that handmade software has added value or quality over slop. I still believe it does right now, but that might not always be true. And this is an industry that has pumped out a lot of sloppy code for decades, even before it was actual slop.
well this is not exactly unique to software. It is not a given that the 'handmade' nature of a product (luxury or otherwise) manifests in anything tangible or self-evidently superior. Luxury products in general overwhelmingly treat the crafted nature of the product as almost solely an investment in narrative (read: marketing) and market positioning, not an actual material outcome.
> you don’t see the quality in handmade code like you do in a high end watch or a luxury automobile or appliance.
But you can't see the "quality" in a high end watch either. They are inferior to low-end watches by every metric you can imagine.
What're you talking about? I can instantly detect if software was created with AI. It has a stench much like a cheap product who's materials off gassing, and Im certainly not using it over the handmade solution. There will always be two options, like in clothes or watches.
Pay increases and red carpet rollouts along with your choice of fruit or candy basket.
Nothing. There will be more of us than ever as the amount of money each engineer can generate goes up. I’m not sure why everyone comes to the conclusion that there will be less engineers as productivity increases. We aren’t that expensive relative to the value we generate and AI, when wielded by an expert, will exponentially increase the amount of return a single engineer can generate.
>I’m not sure why everyone comes to the conclusion that there will be less engineers as productivity increases.
The politics of corporations is such that they don't actually want near infinite productivity. They require churn, cycles, a cadence.
LLMs have taken much of the enjoyment out of coding for me, but I don’t think it has to be that way. I hope that as an industry we settle on LLMs being more like tools than human assistants. I think most of the engineers at my company are using LLMs as human assistants- most of them have agentic workflows set up and have premium subscriptions to Claude AND Gemini AND ChatGPT. Many have local LLMs running on their company MacBook Pros, but they can never manage to describe to me in plain English what those local LLMs are doing. I would compare local LLMs to Raspberry Pi clusters. They’re neat, and technically they can do stuff but they are incredibly impractical. I want so desperately to have the power of Claude Opus running locally, but we are incredibly, ridiculously, extremely far off (benchmarks are often misleading and nobody likes to talk about the paltry tokens/s they’re getting on their home rig).
So the problem for me is that I’ve noticed a trend with my coworkers- they usually don’t have a formal CS background (e.g., Comp Sci degree), their resumes are unimpressive, and they have personal websites where anyone can download a copy of their resume. I haven’t seen this at any previous job. It’s just weird, man. Anyway, so my problem is when they are purported to be the best of the best at this company, and they’re chugging the LLM kool-aid, I’m struggling to figure out whether they’re really ahead of the curve and getting the drop on the rest of us, or they’re doing basically no work and waving through each other’s PRs with little to no oversight on what the LLMs are writing. The bush league nature of many of the bugs I’m seeing going all the way to prod makes me think it’s the latter.
Meanwhile, these guys and their teams are completing an absurd number of story points each sprint. My team is dead last (and coincidentally we’re all pretty anti-AI for writing code).
I’m waiting for the other shoe to drop. I’m waiting for the C suite to call our project a failure and lay everyone off. The product generally works, though I think if the devs were truly rockstars, things would be much smoother. It seems like LLMs are letting middle-of-the-road or mediocre devs appear to be rockstars purely on output speed. Nobody in this org cares if the output has 4x the bugs it should have, just that it was completed quickly.
The main issue I think is that nobody- including the rockstar tech leads- knows how the system works. They don’t know the code. There’s a bug and it takes forever to track down. Don’t ask about unit tests.
I think in a perfect world, these LLMs are treated like another tool rather than another person- meaning you and I still write a large portion of the code, and we read and understand every line- especially the lines generated by LLMs. It’s still a huge productivity multiplier for me to not have to remember how to do mundane things like write out arrays into CSV files- before I’d find a close implementation on StackOverflow and tweak to my needs. Now I can just describe what I want the CSV to look like and Claude Code generally gets it right on the first try.