Biggest problem right now for me is what do I tell my children to pursue in college. Up until last year my default answer is a path towards Software Engineering. Now, I am not even sure anymore.
I can work on this "soul-sucking" job, but what of the next gen?
I should preface my comment with the fact that I’m ultimately neutral on the AI space right now. I’ve gone back and forth on the topic.
That said, I think this article underscores the fact that software engineering as a discipline is already unrecognizable from what it was like even a few years ago. You now have to be more into architecture and directing how the AI development tools write out your code.
I know a guy who works at Amazon, and the models at OpenAI and Anthropic are so good that he doesn’t even really write code in the classic sense anymore. However, because of his background and experience, he does know what he should expect and if an output matches the quality he wants, versus just blindly trusting the AI tool.
In our world, you have to adapt or die. It’s that simple.
What will happen when he retires? What about the generation who grew up pressing enter on claude code, though nowadays auto mode does it for you.
It’s strange for a brief moment in history humans peaked in technological literacy and it went down from there. I hope the doctor guys are still learning.
I have coworkers who report learning a lot about some system by using Claude to do things they would not be able to do on their own, and subsequently seeing more about how the system works.
I suspect that literacy will not go down, but what we become literate in will change. Just as compilers reduced literacy in assembly language, automatic code generation will reduce literacy in code syntax. But then, other forms of literacy will replace that. High level system design, etc.
I think for autodidacts this will hold true and it will be a boon. For those who want an easy way out though it will almost certainly become a mire. Great points, interested to see how it plays out.
Meanwhile, Japanese monks are dismantling shinto temples piece by piece and rebuilding them piece solely so every future generation will know how it’s done.
A few of that generation will be drawn to understand the inner workings, thoroughly, from the very basics and first principles. Because they need to. For inexplicable reasons. They are the George Platen of the 2030s.
There was no economically justifiable reason for me to learn about compilers, operating systems, and even CMOS VLSI. But I couldn't be satisfied writing software without knowing how those things worked. I taped out a chip in which all the polygons drawn by deterministic software I'd eithe written or understood. No place+route. It was the 2000s version of non-AI coding in the 2030s -- standard cell placement is essentially gradient descent.
Thank you for using a reference I didn’t get :). I looked up who George Platen is/was and I’ve now a new thing for my reading list. Also, 100% agree with you
> I know a guy who works at Amazon, and the models at OpenAI and Anthropic are so good that he doesn’t even really write code in the classic sense anymore
"I know a guy", what a weird thing to say in October 2026. I don't think I know anyone who still writes code in the classic sense.
i get paid $300k cash to write all my code by hand. and this is at an AI-pilled company with a CEO pushing AI every day.
boss doesn't know or care. i just get plus perf reviews and raises so it's chill. my perf reviews are actually trending up since opus 4.5 so i'm feeling good actually.
i think smart people crave an adult in the room in these infantile psychotic times
The root of the problem here isnt ai, after all. Bosses who force people to constantly work overtime for free have always existed. And instead of philosophizing about claude the guy should just look for a job with better conditions
Might really be your sector. Claude is massive in my organization, with no one writing code by hand anymore. The biggest struggle is keeping up with what is being done, reviewing and understanding.
What has your code base ended up looking like? Is there a coherent abstraction hierarchy?
I keep checking back on LLMs every couple months and while they are now great for figuring out bugs in a system I don’t know much about (say in my use of a framework I’ve never used before), it’s been super helpful. But whenever I have it write a lot of code, even in a not huge project, it mostly works but it’s like a gum and popsicle stick machine gun and I’m scared of what large org code bases will look like in a few years.
As someone who makes heavy use of AI tools to write code I also find it surprising. The idea that even the latest models of Claude and Codex are capable of writing sufficiently useful and production usable code is baffling. Unless you are operating at extremely low stakes with few meaningful constraints I don't see how anyone can seriously rely fully on AI generated code. I see the code daily, it certainly can save some time when used properly but there are very hard limits to the amount of concepts that can be affirmatively managed before AI begins to tie itself in knots. The only solution AI is capable of from that point is of course generating more code to sort out any issues. "Hand" writing actual clear, concise and unbloated code is still very much a requirement.
Many companies are essentially forcing the use of it via raising the bar for performance by 10x(hyperbole). Yeah it’s not great everywhere and I’m sure there firms that aren’t as pushy.
Wrong company - toxic culture. It's completely possible to use LLMs for development and loving it. Focus on innovation and quality, less on shipping. That's the execs' fault. I'd quit and go somewhere people care about what they're doing.
And I have always found that a little hollow, and depressingly capitalist within tech. Outside of tech, people who nearly exclusively value end-results are the same people taking pictures of their food, and the same people who can't sit in silence, or get irritable when a hike hasn't blessed them with a view yet. For me, it's almost always been the journey, and the result is just an ephemeral blip while work continues iteratively.
Sure, automating away things that would otherwise just be basically copy/pasted anyway have always been fine, but this is a very different scale.
It is interesting how people can differ in terms of disposition though.
Edit: I didn't mean to imply that one disposition is better or worse than another objectively, but rather that I subjectively find impatient, superficial, consumption-driven people extremely dull, which I also tend to associate with people who mostly care about the result. Additionally, I believe that this focus on the production of more and faster is a non-trivial contribution to producing things that have no longevity or are otherwise enshittified. The less we're involved in the intimate details of a thing, the less we're inclined to care, and the less it's even acceptable to care about quality.
I think I understand your point, and it certainly is true if you get in the right mindset with the right anectadata to back it up. But I feel it's leaving a lot on the table. Result oriented people are key to linking the technical with the real world. It's a bit of a wide statement to say that everyone who likes end result is impatient superficial or consumption driver.
I agree that there's a decent overlap here, and that your point stands. But there's a lot of merit in getting real products out there, vibe coded or not. Market forces then dictate value.
I don't find myself getting more work done... I'm getting more help when I get stuck, but a lot of times AI just creates a bigger mess before it solves a problem.
As AI advances, it is natural for development to focus on output and prioritize sprint and PR speed.
So rather than insisting that more review time is needed, teams should set success criteria and a set of reference examples to compare the outcomes against.
As long as there are clear criteria, there is no need to worry about what AI produces; ambiguous parts can simply be left to human judgment.
"ambiguous parts can simply be left to human judgment." - it does not work like that. If a human does not know how the code works, they cannot apply their judgement. And they don't know how the code works if they just press enter all day.
Biggest problem right now for me is what do I tell my children to pursue in college. Up until last year my default answer is a path towards Software Engineering. Now, I am not even sure anymore.
I can work on this "soul-sucking" job, but what of the next gen?
Agricultural science. Science in general. College really shouldn't be solely vocational training.
Why is that a better choice? With respect to career prospects?
Mechatronics is cool
Have hope - go to the htmx site and look at some of Carson’s recent writings. Great stuff in there and he’s a university prof.
> tell my children to pursue...
How about letting them decide? Problem solved for you and them.
I should preface my comment with the fact that I’m ultimately neutral on the AI space right now. I’ve gone back and forth on the topic.
That said, I think this article underscores the fact that software engineering as a discipline is already unrecognizable from what it was like even a few years ago. You now have to be more into architecture and directing how the AI development tools write out your code.
I know a guy who works at Amazon, and the models at OpenAI and Anthropic are so good that he doesn’t even really write code in the classic sense anymore. However, because of his background and experience, he does know what he should expect and if an output matches the quality he wants, versus just blindly trusting the AI tool.
In our world, you have to adapt or die. It’s that simple.
What will happen when he retires? What about the generation who grew up pressing enter on claude code, though nowadays auto mode does it for you. It’s strange for a brief moment in history humans peaked in technological literacy and it went down from there. I hope the doctor guys are still learning.
I have coworkers who report learning a lot about some system by using Claude to do things they would not be able to do on their own, and subsequently seeing more about how the system works.
I suspect that literacy will not go down, but what we become literate in will change. Just as compilers reduced literacy in assembly language, automatic code generation will reduce literacy in code syntax. But then, other forms of literacy will replace that. High level system design, etc.
I think for autodidacts this will hold true and it will be a boon. For those who want an easy way out though it will almost certainly become a mire. Great points, interested to see how it plays out.
Meanwhile, Japanese monks are dismantling shinto temples piece by piece and rebuilding them piece solely so every future generation will know how it’s done.
A few of that generation will be drawn to understand the inner workings, thoroughly, from the very basics and first principles. Because they need to. For inexplicable reasons. They are the George Platen of the 2030s.
There was no economically justifiable reason for me to learn about compilers, operating systems, and even CMOS VLSI. But I couldn't be satisfied writing software without knowing how those things worked. I taped out a chip in which all the polygons drawn by deterministic software I'd eithe written or understood. No place+route. It was the 2000s version of non-AI coding in the 2030s -- standard cell placement is essentially gradient descent.
Some people just have to know how it works.
Thank you for using a reference I didn’t get :). I looked up who George Platen is/was and I’ve now a new thing for my reading list. Also, 100% agree with you
Oh boy[1], do I have[2] articles[3] for[4] you[5].
The history of, 'No one will know how to do anything and the industry will collapse' is almost as old as programming itself.
1 Edsger Dijkstra, "How do we tell truths that might hurt?" (EWD 498, 1975) https://www.cs.utexas.edu/~EWD/transcriptions/EWD04xx/EWD498...
2. John Backus, "The History of FORTRAN I, II, and III" (1978) https://dl.acm.org/doi/10.1145/800025.1198345
3. Dijkstra, "On the cruelty of really teaching computing science" https://www.cs.utexas.edu/~EWD/transcriptions/EWD10xx/EWD103...
4. Alan Perlis, "Epigrams on Programming" (1982) Specifically #93 https://gwern.net/doc/cs/algorithm/1982-perlis.pdf
5. Joel Spolsky, "The Perils of JavaSchools" (2005) https://www.joelonsoftware.com/2005/12/29/the-perils-of-java...
Thank you for sharing. I enjoy reading these.
But don’t you think these were just complaints about higher levels of abstractions? We still had to learn lower levels in school.
AI and tbh society in general just invalidates the economic viability of learning what’s under the foundations.
> I know a guy who works at Amazon, and the models at OpenAI and Anthropic are so good that he doesn’t even really write code in the classic sense anymore
"I know a guy", what a weird thing to say in October 2026. I don't think I know anyone who still writes code in the classic sense.
Is it so surprising that people who work in other domains and specialties may have different experiences than you?
People write code for all sorts of reasons, it's not a monolithic or uniform activity.
i get paid $300k cash to write all my code by hand. and this is at an AI-pilled company with a CEO pushing AI every day.
boss doesn't know or care. i just get plus perf reviews and raises so it's chill. my perf reviews are actually trending up since opus 4.5 so i'm feeling good actually.
i think smart people crave an adult in the room in these infantile psychotic times
The root of the problem here isnt ai, after all. Bosses who force people to constantly work overtime for free have always existed. And instead of philosophizing about claude the guy should just look for a job with better conditions
I still write code, by hand mostly. Perhaps it’s my sector or something, but I’m always quite surprised when I read things like this.
Might really be your sector. Claude is massive in my organization, with no one writing code by hand anymore. The biggest struggle is keeping up with what is being done, reviewing and understanding.
What has your code base ended up looking like? Is there a coherent abstraction hierarchy?
I keep checking back on LLMs every couple months and while they are now great for figuring out bugs in a system I don’t know much about (say in my use of a framework I’ve never used before), it’s been super helpful. But whenever I have it write a lot of code, even in a not huge project, it mostly works but it’s like a gum and popsicle stick machine gun and I’m scared of what large org code bases will look like in a few years.
As someone who makes heavy use of AI tools to write code I also find it surprising. The idea that even the latest models of Claude and Codex are capable of writing sufficiently useful and production usable code is baffling. Unless you are operating at extremely low stakes with few meaningful constraints I don't see how anyone can seriously rely fully on AI generated code. I see the code daily, it certainly can save some time when used properly but there are very hard limits to the amount of concepts that can be affirmatively managed before AI begins to tie itself in knots. The only solution AI is capable of from that point is of course generating more code to sort out any issues. "Hand" writing actual clear, concise and unbloated code is still very much a requirement.
Many companies are essentially forcing the use of it via raising the bar for performance by 10x(hyperbole). Yeah it’s not great everywhere and I’m sure there firms that aren’t as pushy.
10x really isn't hyperbole at this point
Producing 10x the amount of slop isn't being 10x productive.
Certainly just my experience, but my supervisors totally and holistically believe the two are one and the same.
I don’t think any of our developers have written any code since March/April. I wonder if they even know how to anymore.
porting, tests and scaffolding code, documentation is what ai is quite good at and a human is terrible at
you are going to be one of the most valuable software engineers in 5 yrs, dont stop doing what you do!!!
Hard agree. This industry is awful now.
Wrong company - toxic culture. It's completely possible to use LLMs for development and loving it. Focus on innovation and quality, less on shipping. That's the execs' fault. I'd quit and go somewhere people care about what they're doing.
I just started using claude. I am loving it.
For context, I have been programming for 25 years now, mainly out of self interest.
I find joy in relegating mundane tasks to claude, while working on the broader picture with my trusty pen and paper.
I don't give "Build me this end to end" type tasks.
I do the program on paper. The architectural aspects. And instruct claude to flesh out lower level stuff.
I feel the only people struggling with this are people who loved programming and syntax.
Product-minded software engineers are thriving. Less tedium, easier refactors, manageable scope creep.
The “hard” part of engineering has moved up to higher level challenges: evals, sandboxes, etc
Yep. I'm loving this new world. I was always a product minded developer. I loved the end result rather than the code.
And I have always found that a little hollow, and depressingly capitalist within tech. Outside of tech, people who nearly exclusively value end-results are the same people taking pictures of their food, and the same people who can't sit in silence, or get irritable when a hike hasn't blessed them with a view yet. For me, it's almost always been the journey, and the result is just an ephemeral blip while work continues iteratively.
Sure, automating away things that would otherwise just be basically copy/pasted anyway have always been fine, but this is a very different scale.
It is interesting how people can differ in terms of disposition though.
Edit: I didn't mean to imply that one disposition is better or worse than another objectively, but rather that I subjectively find impatient, superficial, consumption-driven people extremely dull, which I also tend to associate with people who mostly care about the result. Additionally, I believe that this focus on the production of more and faster is a non-trivial contribution to producing things that have no longevity or are otherwise enshittified. The less we're involved in the intimate details of a thing, the less we're inclined to care, and the less it's even acceptable to care about quality.
I think I understand your point, and it certainly is true if you get in the right mindset with the right anectadata to back it up. But I feel it's leaving a lot on the table. Result oriented people are key to linking the technical with the real world. It's a bit of a wide statement to say that everyone who likes end result is impatient superficial or consumption driver.
I agree that there's a decent overlap here, and that your point stands. But there's a lot of merit in getting real products out there, vibe coded or not. Market forces then dictate value.
I'm very curious to see how the "building codes are first written in blood" will play out in the software world.
I don't find myself getting more work done... I'm getting more help when I get stuck, but a lot of times AI just creates a bigger mess before it solves a problem.
[dupe] Discussion: https://news.ycombinator.com/item?id=49812975
quit
and do what?
Learn to cod... er... do plumbing?
AKA turd herding
What jobs do non-programmers do?
Barista
[flagged]
[flagged]
[flagged]
As AI advances, it is natural for development to focus on output and prioritize sprint and PR speed.
So rather than insisting that more review time is needed, teams should set success criteria and a set of reference examples to compare the outcomes against.
As long as there are clear criteria, there is no need to worry about what AI produces; ambiguous parts can simply be left to human judgment.
"ambiguous parts can simply be left to human judgment." - it does not work like that. If a human does not know how the code works, they cannot apply their judgement. And they don't know how the code works if they just press enter all day.