People keep giving examples like printed press, or whatever, but honestly the printed press did kill writing/drawing by hand. When was the last time you bought a hand written book or a hand painted picture?
Sure, they still exist, but they are priced and marketed towards a specific class of people, and not the mass. And with software it’s even harder as software was never targeted at the mass. Software was always a by product of a car, a tv, or your fridge. No ordinary folk ever said “I’m going to go a buy WinZip now”. They simply had a variant of WinZip available with whatever computer they bought. There is simply no category of people who are willing to pay for “high quality hand crafted software”, except for perhaps a few ideologically motivated people.
Software development “lost” because it was seen as “something that comes for free with a computer”, and not something that people have to actively buy, like they buy TVs, fridges, or paintings.
The printing press did displace the manual copying of manuscripts, which had insanely low bandwidth and generally kinda sucked for everyone involved. But for several hundred years after, until the industrial revolution and the mass production of mechanical typewriters, pretty much everything printed on a printing press was originally authored by hand. Even with typewriters handwriting lingered for quite a while, since they were very demanding in terms of skill, and all mistakes were permanent.
Arguably it's not until digital writing tools emerged handwriting was truly displaced. That made typing on a keyboard accessible enough that anyone with working hands can do it fairly easily.
One thing that strikes me about these discussions is just how awful the metaphors are.
We're constantly being told that jobs designing and building skyscrapers are close to obsolete because some yokel made a factory that manufactures yurts.
Writing hand crafted code for the sole use of an individual never happened. Writing hand crafted code to be run by millions of people is routine. You can't compare it to a spinning jenny, a printing press, the motor car or indeed any other technological invention that scaled production. It is nonsensical by definition.
I think a lot of people just dont want to hear that LLMs are a step change and not actually revolutionary. Theres too much invested in them being world changing at this point. Blockchain was the same.
Also hard to overstate the sheer impact of a tool like Excel, which in many ways is an ad-hoc programming tool so intuitive that anyone can learn to use it. Even though it sees more jeer than celebration, and is often something techies seek to liberate users from, it's likely one of the most economically impactful inventions of the 20th century, already delivering many of the things we hype AI for.
Factories making canned or microwave meals killed chefs. Why would anyone cook food themselves if they can get food ready made only missing heating. Cooking and being chef is dead. Factory produced slop is the new future...
Printing press did not kill writing at all, only the job of scribes. One still had to write their ideas down and work on a manuscript and some one had to set the press up (not trivial at all). All that is to say that I think the ability to write code from scratch is still necessary while I don't really think the printing press metaphor is apt I think being able to write scripts and write small programs to do stuff you want by hand is still useful even though it's true that LLM's can do this sometimes you want to understand the thing fully and to work exactly how you want it.
I do not think programming is dead, but because I don't think LLMs can really do the job unassisted. I agree with you that the printing press is a dumb argument, basically tantamount to saying programming is dead except for bespoke artisanal programs to buy at ren faires and hacker conferences.
It accepts the inevitability premise so completely that I wonder if the author is secretly an AI booster trying to make the other side look bad.
> When was the last time you bought a hand written book or a hand painted picture
So programming "ended" if it's no longer such a dominant profession, even if plenty of people still do it as an art/craft/hobby and some of them even get paid to do so?
Doesn't sound like the "end of programming" to me, more a reduction in the number of programming jobs. One that if anything is overdue - many developers today get paid to do jobs that shouldn't need to exist even without AI.
> And with software it’s even harder as software was never targeted at the mass.
Do products like Adobe Photoshop or Microsoft Office Suite not count as “targeted at mass?” People—normal people—absolutely used to pay for software before SaaS took hold.
I don’t think the problem is that normal people don’t want to buy ““artisan software”” but that people are so cynical about software companies that giving them any money at all is asking to have the rug pulled out from under you.
The printing press killed the work of scribes, people who copied books page by page. It was a form of art and craftsmanship. It didn't disappear completely, but 99% of production moved to the printing press.
Programming will follow the same path. Typing code by hand will be relegated to 1%, while 99% will be generated by assistants and agents.
I’m confused by your example. Prior to the printing press, hand written books (or any reading material) would not have been accessible to the masses; hence the low literacy rates.
Furthermore, while no one was saying “I’m going to go buy WinZip now”, many were lining up to buy WinRAR?
Finally, the issue with “high quality hand crafted software” is that software and similar projects are made possible through engineering not through artistry.
Simply put:
"Engineering is achieving function while avoiding failure." — Henry Petroski
People really need to stop finding one vaguely similar situation in the past and say "this is what will happen". The past is a different country, and the outcomes depend on many factors and aren't predestined.
I suppose the author is pointing at the fact that coding != programming. The term "programming" is wider applicable than just in relation with computers.
Though I don't quite agree with the analogy. To me it seems more like the difference between writing with pen & paper or typing, digital vs analog drawing or printing press vs photo copier. There are some significant changes in how to use the tool properly (can't backspace/undo on physical). However the essence of writing a good story is not in the typing/penning it transcends that. At the same time it is true that different individuals may produce better output with one or the other due to personal preferences or adapting a different creative process.
Every other time that programming has gotten easier we got more (often worse) code, though some folks who made their livings exploiting the inefficiencies had to adapt. This will be the same.
This isn't the end of programming, it's the beginning of your broccoli haired nephews ability to program. It's also opening a whole new world of opportunity to those with real skill.
A world filled with mountains of terrible code is a world that will reward the very programers who are trying to fight that future today.
Yeah right, like I want to take over the shit broccoli haired nephews created... Even for 3x paycheck.. The one who takes it is a looser, being exploited for pennies to fix all the shit. And I bet at the end, you will be fired anyway, because you are not of their kind..
Winners are managers (who hired nephews) and nephews themselfs, they did barely any usefull job and got paid a lot. They should do usefull job instead, like drivers, warehouse jobs, services, farming and other less paid but very importand and usefull jobs.
> Winners are managers (who hired nephews) and nephews themselfs, they did barely any usefull job and got paid a lot.
That is what the hype is about. They created their kingdom/empire with software and data. And they’ve been looking for any solution that promises to replace programmers. LLM is the best mirage ao far and have grown to cult level.
> This isn't the end of programming, it's the beginning of your broccoli haired nephews ability to program
Which is funny because that was the case 25+ years ago. We (the nephews) would create websites using PHP. It was common for 15, 16 years old teenagers to know more about tech than the adults. That's not the case anymore. So this time it might be different.
If someone comes with a pretty decent idea and execution, it doesn’t really matter if code was written by human or llm, does it? Like if the next torvalds ships something great in the open, who cares how was it written.
> it doesn’t really matter if code was written by human or llm
In theory no, but for a successful execution you must have people that understand what's going on, otherwise the competitor that understands will eat you.
> who cares how was it written.
In the pre AI era a lot of sw was written by interns, juniors, some bigcon consultants, some offshore indian or east-european dudes, it didn't even matter before.
One thing I keep discovering about the agentic era of programming is that all of the old lessons and bits of wisdom about programming are still valid. They're just 10 to 100x more important now that you can generate 10,000 lines in an hour.
e.g. Break your code up into loosely coupled modules, don't have leaky abstractions, don't have huge God files, have good black box end-to-end testing, do the design work before the implementation, do lots of automated testing, have good documentation, etc.
Here is another one that was always true, end users only care about the observable behavior of your program. Does it solve the problem in a performant way? Is it reliable?
And another one, it was common when you inherited a large code base that was mature and written before you to not have any clue how most of it worked.
We have to accept that the art of nudging the text around is just not part of producing code for money anymore.
In my experience little of what you said is important is actually needed:
> Break your code up into loosely coupled modules
> don't have leaky abstractions
Not needed. The AI handles breaking up code and I mostly don't review abstractions.
> don't have huge God files
Less true than ever. The AI can read a 10k line file the same way it can read 20 files with 500 lines.
> have good black box end-to-end testing
The AI does the testing by default, and it is fine. I don't write the docs nor do I do design work.
Now what I said is certainly going to offend some users here, but that's how it is, and I know the quality of the product built with this workflow.
This workflow works well for software you develop alone, it is going to be trickier to adopt it in a team.
Still, in my opinion the writing is on the wall for our profession. No one can say with any certainty to what degree our expertise is still needed in the software development process in a year or two.
Agree. Now that we're seeing the quality of software we use worsening (not improving!) with increased usage of LLMs to produce code, it's even more important than before to get the fundamentals solid. I don't mean "fundamentals" as in "easy things in CS", but the mechanisms behind the systems we build (on top of). And, not being fresh out of school and having experience _suffering_ with poorly designed systems / processes is also still an advantage, as that suffering will ultimately let one make more informed design choices.
> Like if the next torvalds ships something great in the open, who cares how was it written.
I was reading a biography of Tovalds and one of the core things that mark the birth of Linux was the drive to learn. Same when reading “Coders at Work”, the will to learn and make things happen is a common recurrence.
I’ve never seen LLM coding that is not “I can’t bother to learn about this”. It’s always about magicking answers out of thin air, not about motivation to explore some fields.
I have resigned to the fact that there be dragons.
I can tell you how things are with Sol.
With Sol you need to know what you're doing and roughly how you'd implement it by hand yourself in order to use Sol effectively. With Sol you need to look over it's work and generally treat it like a junior (reviewing critical sections yourself, reviewing throwaway code and prototypes only cursorily). With Sol it's usually not profitable to write code yourself. With Sol it's basically never profitable to debug code yourself. With Sol you cannot trust it to auto-deploy stuff without an entire pipeline and guardrails.
I have no idea how you will have to work with what comes next.
I think that coding by hand is largely not going to occur anymore. What that does to the learning of new developers + the stagnation of old developers? I do not know.
I believe that what makes a good developer (a great developer, really) is the ability to do something at the smallest level "perfectly" or close to it. Chaining those tiny building blocks together. Understanding the entire flow of data + logically. Making something as fast as it can be. When you put more and more of those tiny blocks together, things get a little more imperfect - but a fantastic developer can keep them all perfect for a long time.
We're jumping way above that level. That small, perfect logic process, that visibility is now completely hidden from us, some of us wouldn't even know how to do it anymore. And juniors of today will never learn to.
And maybe they don't need to. Perhaps we will just build "bigger" software so that the level of imperfection becomes less noticeable. Larger systems, more functionality, more security.
I think the brain atrophy is going to hit us HARD though. It feels too fast to rip away the one thing that taught us all of this. Building something piece by piece was the way most of us learned. Then you can jump into bigger and bigger projects and your brain creates those pathways. Previous generations had months / years to learn new tooling - and now we threw ALL of the tooling away and said "use this instead". I think good developers will always find a way to learn and grow, but man is it going to take some time, we completely shifted the way the entire industry produces the machine in the matter of 12 months.
I still believe that code is the best way to represent a sufficiently complex specification, natural human language feels very ambiguous sometimes, like extra lossy compression (whatever that is worth).
It's really interesting to the vagueness of natural language and the fact that it's actually impossible to express a single concrete meaning in natural language. That is to say their is always a set of possible interpretations one can never simple refer to one single thing and a program in a programming language is always only one thing (I know that you could argue that compiling to different machine code means their is vagueness since it might produce a x86-64 binary or a powerpc binary and the issue of UB but you see my point).
I do think that if one simply 'vibe codes' something they can never really understand the program even if they prompted it unless they study the source in detail. In natural language you can only ever point to a set of things that containing a subset of thing(s) you want.
I write code in my work, and enjoy it. I also write less directly because of llm's but get to read more (which I luckily also enjoy, as long as I understand what is going on). I think you made more of a snide comment and deep down you actually enjoy it ;-). I mean, who doesn't like to code. It's awesome. If not, i hope you find something fulfilling sooner rather than later.
I love to write code and solve problems in personal projects. I love to read interesting technical books. I love to learn certain new technologies.
I hate doing the same at work. I couldn’t care less about my CEO’s product, nor about our linting tips or dependency management rules. My PO can pretend he’s the owner of the company, my engineer colleagues that we are shipping the next google. I couldn’t care less. I do it because I get paid and because it’s my career.
So, it’s not like “you like coding or not”. It has nuances
We've seen this cycle before with high-level languages or IDEs, where the bar for entry drops but the underlying complexity of building reliable systems remains the same.
The responses about code quality, wisdom that comes with experience, etc. are missing the big picture: no one will ever care about any of this as long as things work as explained in the prompt.
Eventual shortcomings will be fixed not by looking into the code, but by typing in another prompt, asking another agent.
Doing it manually will not bring any value to anyone.
I saw it already with my own eyes: a friend of mine started an online psychology counseling site. A year ago they would hire a developer. Now, he only needed an LLM. And it’s the whole thing: a booking system, video calls, a website. What’s underneath: he doesn’t know. The only thing that matters to him is that it works.
Can he trust it's secure? Does what about when something goes wrong. Ops issues might come into play and just handing AI agent ssh keys could cause all sorts of issues. Only time will tell though I guess.
Gotchas aside: I’m not saying the end of programming is a good thing. I’m just saying: it’s happening. Unless the LLM tech will suddenly hit any sort of ceiling.
I feel like the article in question was actually saying some like others are expressing here that writing code by hand is over not the other aspects despite the title.
I also think the metaphor of drawing is not perfect but a quite good in this comment in that drawing is talking about something that is missed on lots of analysis, drawing is a free form kind of sketching and one can quickly draw something and work things out. I think being able to quickly code something by hand is still valuable and though you can get an llm to write a script sometimes you just want to write it yourself so you know the ins and out of your 100 cloc script.
I find it interesting just how different the responses to this new tech are.
I wonder if there is an inverse correlation between "usage of local models" and "despair", or if I am just making that up.
It would make sense, I suppose, given that local models still come with quite the upfront cost. So a lot of people are experiencing LLMs purely through the user-hostile cloud.
the technology isnt the problem at all. it's just being used as an thin pretext to devalue and treat devs like shit by the followers of a new executive religion.
Typing code by hand is dead I agree. But calling it the end of programming was bit weird though.
I use to write a lot of code everyday for more than 2 decades, but now with LLMs, I can get most of my code written by LLMs, and I make sure they write the code that I want to. Same as, when I had to manage a team of juniors.
Many interesting metaphors to find, "it's like the printing press", "it's like IDE's", "it's like going from hardwired programming to assembly language". At the same time it's unlike anything we've seen before. Interesting to see where the journey goes, I am along for the ride.
Script kiddies aren't really a problem, but there's a place where technically competent CTO types can sooner/inexpensively solve problems with AI rather than documenting what they want done.
>It's not the end of programming, in the same way as the invention of the printing press wasn't "the end of drawing by hand".
The printing press let you scale copies of something already drawn. That is nothing like what AI is doing here which is making drawing easier. It's more like people now draw by hand using pens and pencils instead of scraping small rocks against bigger rocks. Technically you can still scrape rocks together, but everyone opts for the modern convenience of better tools.
>It's not the end of programming, in the same way as the birth of big tech wasn't the beginning of programming.
Considering big tech brought programming to the masses arguable big tech did begin programming as a thing people do.
I think AI brings in many problems. It also brings in new functionality for people unable or unwilling to write working programs on their own.
My number #1 gripe of AI is that you become dependent on external companies. I am no loner willing to go that route in general, but even less so when it comes to greedy US corporations that constantly abuse mankind.
My number #2 gripe is that AI is ... dumb. I am noticing this on youtube. So many fake-videos appearing, and while I in general do not watch them for the most part, they are incredibly deceitful. And ... dumb. It's just a waste of time to watch them, even more so than it is already a waste of time watching videos on youtube (I often have them just for background music in the browser, for music at the least).
AI also makes projects less interesting to contribute to. If 99% of the contributions come by AI, why would I want to be bothered to "get involved"? Ruby has this problem right now. Granted, AI is a helper tool, but boy, what a fake it is - Hiroshi Shibata suddenly has perfect english skills, but it is AI slop generated. So I am no longer "interacting" with humans but with the AI slop. That's not worth my time.
AI also is very ruthless against programming languages being bad in certain areas. Case in point: documentation. If AI can slop AI spam generate documentation, even if it is not perfect, whereas I need to manually search for ruby documentation while Google has crippled its search engine to become useless, then this is now a huge disadvantage I am at with. Searching for stuff takes time. Ruby is just an example here; you can extend this to many other projects, including non-programming languages. We need better documentation, but oddly enough the documentation is constantly getting worse in the last +5 years or so. AI skynet is cannibalizing us - people don't seem to realise or care. I somewhat gave up on trying to nudge people to seriously improve their documentation, they just don't seem to want to (not all, but most).
People keep giving examples like printed press, or whatever, but honestly the printed press did kill writing/drawing by hand. When was the last time you bought a hand written book or a hand painted picture?
Sure, they still exist, but they are priced and marketed towards a specific class of people, and not the mass. And with software it’s even harder as software was never targeted at the mass. Software was always a by product of a car, a tv, or your fridge. No ordinary folk ever said “I’m going to go a buy WinZip now”. They simply had a variant of WinZip available with whatever computer they bought. There is simply no category of people who are willing to pay for “high quality hand crafted software”, except for perhaps a few ideologically motivated people.
Software development “lost” because it was seen as “something that comes for free with a computer”, and not something that people have to actively buy, like they buy TVs, fridges, or paintings.
Printing presses are not typewriters.
The printing press did displace the manual copying of manuscripts, which had insanely low bandwidth and generally kinda sucked for everyone involved. But for several hundred years after, until the industrial revolution and the mass production of mechanical typewriters, pretty much everything printed on a printing press was originally authored by hand. Even with typewriters handwriting lingered for quite a while, since they were very demanding in terms of skill, and all mistakes were permanent.
Arguably it's not until digital writing tools emerged handwriting was truly displaced. That made typing on a keyboard accessible enough that anyone with working hands can do it fairly easily.
One thing that strikes me about these discussions is just how awful the metaphors are.
We're constantly being told that jobs designing and building skyscrapers are close to obsolete because some yokel made a factory that manufactures yurts.
Writing hand crafted code for the sole use of an individual never happened. Writing hand crafted code to be run by millions of people is routine. You can't compare it to a spinning jenny, a printing press, the motor car or indeed any other technological invention that scaled production. It is nonsensical by definition.
I think a lot of people just dont want to hear that LLMs are a step change and not actually revolutionary. Theres too much invested in them being world changing at this point. Blockchain was the same.
> Writing hand crafted code for the sole use of an individual never happened.
It seems bizarre in a way to see this sentiment on hacker news.
as an economic activity. i made my own crappy todo apps for fun too but I never once got paid for it.
now that I can get a machine to do that instead has the world fundamentally changed?
it is still wrong. Lots of people created and still create software that just themselves use as competitive advantage in their area of work.
Also hard to overstate the sheer impact of a tool like Excel, which in many ways is an ad-hoc programming tool so intuitive that anyone can learn to use it. Even though it sees more jeer than celebration, and is often something techies seek to liberate users from, it's likely one of the most economically impactful inventions of the 20th century, already delivering many of the things we hype AI for.
Entire businesses run off Excel spreadsheets.
Factories making canned or microwave meals killed chefs. Why would anyone cook food themselves if they can get food ready made only missing heating. Cooking and being chef is dead. Factory produced slop is the new future...
Printing press did not kill writing at all, only the job of scribes. One still had to write their ideas down and work on a manuscript and some one had to set the press up (not trivial at all). All that is to say that I think the ability to write code from scratch is still necessary while I don't really think the printing press metaphor is apt I think being able to write scripts and write small programs to do stuff you want by hand is still useful even though it's true that LLM's can do this sometimes you want to understand the thing fully and to work exactly how you want it.
Drawing, not buying hand-painted pictures. Writing, not buying hand-written books. You seem to have fundamentally misunderstood the argument.
I do not think programming is dead, but because I don't think LLMs can really do the job unassisted. I agree with you that the printing press is a dumb argument, basically tantamount to saying programming is dead except for bespoke artisanal programs to buy at ren faires and hacker conferences.
It accepts the inevitability premise so completely that I wonder if the author is secretly an AI booster trying to make the other side look bad.
> When was the last time you bought a hand written book or a hand painted picture
So programming "ended" if it's no longer such a dominant profession, even if plenty of people still do it as an art/craft/hobby and some of them even get paid to do so?
Doesn't sound like the "end of programming" to me, more a reduction in the number of programming jobs. One that if anything is overdue - many developers today get paid to do jobs that shouldn't need to exist even without AI.
> And with software it’s even harder as software was never targeted at the mass.
Do products like Adobe Photoshop or Microsoft Office Suite not count as “targeted at mass?” People—normal people—absolutely used to pay for software before SaaS took hold.
I don’t think the problem is that normal people don’t want to buy ““artisan software”” but that people are so cynical about software companies that giving them any money at all is asking to have the rug pulled out from under you.
The printing press killed the work of scribes, people who copied books page by page. It was a form of art and craftsmanship. It didn't disappear completely, but 99% of production moved to the printing press.
Programming will follow the same path. Typing code by hand will be relegated to 1%, while 99% will be generated by assistants and agents.
Is this a bot comment? You’re just repeating the comment you replied to.
No, it's not a bot. I just tried to answer the same deal in different comments.
> yping code by hand will be relegated to 1%, while 99% will be generated by assistants and agents.
The 1% will be 100x bigger than the 100% of 10 years ago.
As well, but the 99% will be how much bigger? Well, last week GitHub incident shown the commits and PR grown by 14 times.
>When was the last time you bought a hand written book or a hand painted picture?
comparing scribes to writers and artists and software developers is insane
I’m confused by your example. Prior to the printing press, hand written books (or any reading material) would not have been accessible to the masses; hence the low literacy rates.
Furthermore, while no one was saying “I’m going to go buy WinZip now”, many were lining up to buy WinRAR?
Finally, the issue with “high quality hand crafted software” is that software and similar projects are made possible through engineering not through artistry.
Simply put:
"Engineering is achieving function while avoiding failure." — Henry Petroski
Not all software is (or should be) engineering…
People really need to stop finding one vaguely similar situation in the past and say "this is what will happen". The past is a different country, and the outcomes depend on many factors and aren't predestined.
I suppose the author is pointing at the fact that coding != programming. The term "programming" is wider applicable than just in relation with computers.
Though I don't quite agree with the analogy. To me it seems more like the difference between writing with pen & paper or typing, digital vs analog drawing or printing press vs photo copier. There are some significant changes in how to use the tool properly (can't backspace/undo on physical). However the essence of writing a good story is not in the typing/penning it transcends that. At the same time it is true that different individuals may produce better output with one or the other due to personal preferences or adapting a different creative process.
Every other time that programming has gotten easier we got more (often worse) code, though some folks who made their livings exploiting the inefficiencies had to adapt. This will be the same.
This isn't the end of programming, it's the beginning of your broccoli haired nephews ability to program. It's also opening a whole new world of opportunity to those with real skill.
A world filled with mountains of terrible code is a world that will reward the very programers who are trying to fight that future today.
Yeah right, like I want to take over the shit broccoli haired nephews created... Even for 3x paycheck.. The one who takes it is a looser, being exploited for pennies to fix all the shit. And I bet at the end, you will be fired anyway, because you are not of their kind..
Winners are managers (who hired nephews) and nephews themselfs, they did barely any usefull job and got paid a lot. They should do usefull job instead, like drivers, warehouse jobs, services, farming and other less paid but very importand and usefull jobs.
> Winners are managers (who hired nephews) and nephews themselfs, they did barely any usefull job and got paid a lot.
That is what the hype is about. They created their kingdom/empire with software and data. And they’ve been looking for any solution that promises to replace programmers. LLM is the best mirage ao far and have grown to cult level.
What value does the broccoli haired nephew provide in that scenario?
endless job opportunities for millenials and older
> This isn't the end of programming, it's the beginning of your broccoli haired nephews ability to program
Which is funny because that was the case 25+ years ago. We (the nephews) would create websites using PHP. It was common for 15, 16 years old teenagers to know more about tech than the adults. That's not the case anymore. So this time it might be different.
Luckily the kids this time have lots of different opportunities than just tech, they have crypto, social media, and etc
I can't tell if this is ironic or not.
If someone comes with a pretty decent idea and execution, it doesn’t really matter if code was written by human or llm, does it? Like if the next torvalds ships something great in the open, who cares how was it written.
> it doesn’t really matter if code was written by human or llm
In theory no, but for a successful execution you must have people that understand what's going on, otherwise the competitor that understands will eat you.
> who cares how was it written.
In the pre AI era a lot of sw was written by interns, juniors, some bigcon consultants, some offshore indian or east-european dudes, it didn't even matter before.
> who cares how was it written.
Everyone who wants it to work properly.
One thing I keep discovering about the agentic era of programming is that all of the old lessons and bits of wisdom about programming are still valid. They're just 10 to 100x more important now that you can generate 10,000 lines in an hour.
e.g. Break your code up into loosely coupled modules, don't have leaky abstractions, don't have huge God files, have good black box end-to-end testing, do the design work before the implementation, do lots of automated testing, have good documentation, etc.
Here is another one that was always true, end users only care about the observable behavior of your program. Does it solve the problem in a performant way? Is it reliable?
And another one, it was common when you inherited a large code base that was mature and written before you to not have any clue how most of it worked.
We have to accept that the art of nudging the text around is just not part of producing code for money anymore.
In my experience little of what you said is important is actually needed:
> Break your code up into loosely coupled modules
> don't have leaky abstractions
Not needed. The AI handles breaking up code and I mostly don't review abstractions.
> don't have huge God files
Less true than ever. The AI can read a 10k line file the same way it can read 20 files with 500 lines.
> have good black box end-to-end testing
The AI does the testing by default, and it is fine. I don't write the docs nor do I do design work.
Now what I said is certainly going to offend some users here, but that's how it is, and I know the quality of the product built with this workflow.
This workflow works well for software you develop alone, it is going to be trickier to adopt it in a team.
Still, in my opinion the writing is on the wall for our profession. No one can say with any certainty to what degree our expertise is still needed in the software development process in a year or two.
Agree. Now that we're seeing the quality of software we use worsening (not improving!) with increased usage of LLMs to produce code, it's even more important than before to get the fundamentals solid. I don't mean "fundamentals" as in "easy things in CS", but the mechanisms behind the systems we build (on top of). And, not being fresh out of school and having experience _suffering_ with poorly designed systems / processes is also still an advantage, as that suffering will ultimately let one make more informed design choices.
Or do we want to be seeing that?
> Like if the next torvalds ships something great in the open, who cares how was it written.
I was reading a biography of Tovalds and one of the core things that mark the birth of Linux was the drive to learn. Same when reading “Coders at Work”, the will to learn and make things happen is a common recurrence.
I’ve never seen LLM coding that is not “I can’t bother to learn about this”. It’s always about magicking answers out of thin air, not about motivation to explore some fields.
>Like if the next torvalds ships something great in the open, who cares how was it written.
nope. but, vibe coded slop is all we're actually seeing.
I have resigned to the fact that there be dragons.
I can tell you how things are with Sol.
With Sol you need to know what you're doing and roughly how you'd implement it by hand yourself in order to use Sol effectively. With Sol you need to look over it's work and generally treat it like a junior (reviewing critical sections yourself, reviewing throwaway code and prototypes only cursorily). With Sol it's usually not profitable to write code yourself. With Sol it's basically never profitable to debug code yourself. With Sol you cannot trust it to auto-deploy stuff without an entire pipeline and guardrails.
I have no idea how you will have to work with what comes next.
I think that coding by hand is largely not going to occur anymore. What that does to the learning of new developers + the stagnation of old developers? I do not know.
I believe that what makes a good developer (a great developer, really) is the ability to do something at the smallest level "perfectly" or close to it. Chaining those tiny building blocks together. Understanding the entire flow of data + logically. Making something as fast as it can be. When you put more and more of those tiny blocks together, things get a little more imperfect - but a fantastic developer can keep them all perfect for a long time.
We're jumping way above that level. That small, perfect logic process, that visibility is now completely hidden from us, some of us wouldn't even know how to do it anymore. And juniors of today will never learn to.
And maybe they don't need to. Perhaps we will just build "bigger" software so that the level of imperfection becomes less noticeable. Larger systems, more functionality, more security.
I think the brain atrophy is going to hit us HARD though. It feels too fast to rip away the one thing that taught us all of this. Building something piece by piece was the way most of us learned. Then you can jump into bigger and bigger projects and your brain creates those pathways. Previous generations had months / years to learn new tooling - and now we threw ALL of the tooling away and said "use this instead". I think good developers will always find a way to learn and grow, but man is it going to take some time, we completely shifted the way the entire industry produces the machine in the matter of 12 months.
I still believe that code is the best way to represent a sufficiently complex specification, natural human language feels very ambiguous sometimes, like extra lossy compression (whatever that is worth).
It's really interesting to the vagueness of natural language and the fact that it's actually impossible to express a single concrete meaning in natural language. That is to say their is always a set of possible interpretations one can never simple refer to one single thing and a program in a programming language is always only one thing (I know that you could argue that compiling to different machine code means their is vagueness since it might produce a x86-64 binary or a powerpc binary and the issue of UB but you see my point).
I do think that if one simply 'vibe codes' something they can never really understand the program even if they prompted it unless they study the source in detail. In natural language you can only ever point to a set of things that containing a subset of thing(s) you want.
Well I can't wait until I don't have to hand code at work anymore. I also can't wait until I do something else for work.
Become a manager? retire?
I write code in my work, and enjoy it. I also write less directly because of llm's but get to read more (which I luckily also enjoy, as long as I understand what is going on). I think you made more of a snide comment and deep down you actually enjoy it ;-). I mean, who doesn't like to code. It's awesome. If not, i hope you find something fulfilling sooner rather than later.
I love to write code and solve problems in personal projects. I love to read interesting technical books. I love to learn certain new technologies.
I hate doing the same at work. I couldn’t care less about my CEO’s product, nor about our linting tips or dependency management rules. My PO can pretend he’s the owner of the company, my engineer colleagues that we are shipping the next google. I couldn’t care less. I do it because I get paid and because it’s my career.
So, it’s not like “you like coding or not”. It has nuances
Yes! Personally, I love writing software, but I would never want to do it at a company.
You don’t have to wait! Quit and go do something else!
What would you be interested doing?
We've seen this cycle before with high-level languages or IDEs, where the bar for entry drops but the underlying complexity of building reliable systems remains the same.
The responses about code quality, wisdom that comes with experience, etc. are missing the big picture: no one will ever care about any of this as long as things work as explained in the prompt.
Eventual shortcomings will be fixed not by looking into the code, but by typing in another prompt, asking another agent.
Doing it manually will not bring any value to anyone.
I saw it already with my own eyes: a friend of mine started an online psychology counseling site. A year ago they would hire a developer. Now, he only needed an LLM. And it’s the whole thing: a booking system, video calls, a website. What’s underneath: he doesn’t know. The only thing that matters to him is that it works.
That to me looks like the end of programming.
Can he trust it's secure? Does what about when something goes wrong. Ops issues might come into play and just handing AI agent ssh keys could cause all sorts of issues. Only time will tell though I guess.
> Eventual shortcomings will be fixed not by looking into the code, but by typing in another prompt, asking another agent.
So, you'll rely on the tech that created buggy code to somehow suddenly create bug-free code?
Isn’t it the same as relying on human developers?
Gotchas aside: I’m not saying the end of programming is a good thing. I’m just saying: it’s happening. Unless the LLM tech will suddenly hit any sort of ceiling.
Programming survived assemblers, compilers, and frameworks. AI is just the next layer up.
I feel like the article in question was actually saying some like others are expressing here that writing code by hand is over not the other aspects despite the title.
I also think the metaphor of drawing is not perfect but a quite good in this comment in that drawing is talking about something that is missed on lots of analysis, drawing is a free form kind of sketching and one can quickly draw something and work things out. I think being able to quickly code something by hand is still valuable and though you can get an llm to write a script sometimes you just want to write it yourself so you know the ins and out of your 100 cloc script.
> but I remember a time when we decided when a given technology began and when did it end
…when was that?
I find it interesting just how different the responses to this new tech are.
I wonder if there is an inverse correlation between "usage of local models" and "despair", or if I am just making that up.
It would make sense, I suppose, given that local models still come with quite the upfront cost. So a lot of people are experiencing LLMs purely through the user-hostile cloud.
the technology isnt the problem at all. it's just being used as an thin pretext to devalue and treat devs like shit by the followers of a new executive religion.
Typing code by hand is dead I agree. But calling it the end of programming was bit weird though. I use to write a lot of code everyday for more than 2 decades, but now with LLMs, I can get most of my code written by LLMs, and I make sure they write the code that I want to. Same as, when I had to manage a team of juniors.
Many interesting metaphors to find, "it's like the printing press", "it's like IDE's", "it's like going from hardwired programming to assembly language". At the same time it's unlike anything we've seen before. Interesting to see where the journey goes, I am along for the ride.
Script kiddies aren't really a problem, but there's a place where technically competent CTO types can sooner/inexpensively solve problems with AI rather than documenting what they want done.
The best programmers will use AI to handle tedious parts and focus on the hard problems.
Every time tools got easier we just wrote more software, not less. This will be the same.
And it's probably going to be a lot more inefficient. See the rise of electron-based software.
We a have a decade of Government software to clean up, so we are good for a while
And the bot will dump on us another decade's worth - just 10x quicker.
programming has changed and leveled up due to ai it's as big as going from zeros and ones to assembly language
I think the shift is bigger.
In the 01->asm revolution you needed people with roughly the same kind of talent, background and education. Just a different specialization.
Now, writing software is available to people who would have never been able to learn it before.
Programming, much like artisinal chairbuilding, will be demoted to a hobby or niche.
Software engineering will remain.
The fact that we had "The End of Programming" then "No one wants artisanal code anymore" and now "It's not the end of programming"
Getting a really really artificial vibe about the content being pushed to this website :/
Making me feel less motivated to even come here to be honest
Related:
The End of Programming
https://news.ycombinator.com/item?id=49445717
>It's not the end of programming, in the same way as the invention of the printing press wasn't "the end of drawing by hand".
The printing press let you scale copies of something already drawn. That is nothing like what AI is doing here which is making drawing easier. It's more like people now draw by hand using pens and pencils instead of scraping small rocks against bigger rocks. Technically you can still scrape rocks together, but everyone opts for the modern convenience of better tools.
>It's not the end of programming, in the same way as the birth of big tech wasn't the beginning of programming.
Considering big tech brought programming to the masses arguable big tech did begin programming as a thing people do.
Fair, but won't big tech collapse without the actual people force?
No one suggested that a people force would go away.
It’s the end of programming in non natural language
I think AI brings in many problems. It also brings in new functionality for people unable or unwilling to write working programs on their own.
My number #1 gripe of AI is that you become dependent on external companies. I am no loner willing to go that route in general, but even less so when it comes to greedy US corporations that constantly abuse mankind.
My number #2 gripe is that AI is ... dumb. I am noticing this on youtube. So many fake-videos appearing, and while I in general do not watch them for the most part, they are incredibly deceitful. And ... dumb. It's just a waste of time to watch them, even more so than it is already a waste of time watching videos on youtube (I often have them just for background music in the browser, for music at the least).
AI also makes projects less interesting to contribute to. If 99% of the contributions come by AI, why would I want to be bothered to "get involved"? Ruby has this problem right now. Granted, AI is a helper tool, but boy, what a fake it is - Hiroshi Shibata suddenly has perfect english skills, but it is AI slop generated. So I am no longer "interacting" with humans but with the AI slop. That's not worth my time.
AI also is very ruthless against programming languages being bad in certain areas. Case in point: documentation. If AI can slop AI spam generate documentation, even if it is not perfect, whereas I need to manually search for ruby documentation while Google has crippled its search engine to become useless, then this is now a huge disadvantage I am at with. Searching for stuff takes time. Ruby is just an example here; you can extend this to many other projects, including non-programming languages. We need better documentation, but oddly enough the documentation is constantly getting worse in the last +5 years or so. AI skynet is cannibalizing us - people don't seem to realise or care. I somewhat gave up on trying to nudge people to seriously improve their documentation, they just don't seem to want to (not all, but most).
Now we get to program people (and program ourselves) so the business processes are optimally interacting with the ultimate state machine of society
[dead]