Programming is morphing from a creative craft to a dismal science

[pyeri]

To be fair, it had already started happening much before AI came when programmer roles started getting commoditized into "Python coder", "PHP scripter", "dotnet developer", etc. Though these exact phrases weren't used in job descriptions, this is how recruiters and clients started referring programmers as such.

But LLMs took it a notch even further, coders have started morphing into LLM prompters today, that is primarily how software is getting produced. They still must baby sit these LLMs presently, reviewing and testing the code thoroughly before pushing it to the repo for CI/CD. A few more years and even that may not be needed as the more enhanced LLM capabilities like "reasoning", "context determination", "illumination", etc. (maybe even "engineering"!) would have become part of gpt-9 or whatever hottest flavor of LLM be at that time.

The problem is that even though the end result would be a very robust running program that reeks of creativity, there won't be any human creativity in that. The phrase dismal science was first used in reference to economics by medieval scholars like Thomas Carlyle. We can only guess their motivations for using that term but maybe people of that time thought that economics was somehow taking away the life force from society of humans, much similar to the way many feel about AI/LLM today?

Now I understand the need for putting food on the table. To survive this cut throat IT job market, we must adapt to changing trends and technologies and that includes getting skilled with LLM. Nonetheless, I can't help but get a very dismal feeling about this new way of software development, don't you?

[WendyDolan]

I have noticed this too in some jobs—lots of rules and frameworks. Does anyone feel like hobby projects are still creative?

[OrioneSolutions]

That’s such an interesting perspective, and I totally get where you’re coming from. It does feel like programming has shifted from being a creative process to something more mechanical and automated lately. But I think there’s still room for creativity — just in a different way. Instead of writing every line of code, developers now need to think more strategically: how to design systems, prompt effectively, and ensure ethical, high-quality outputs. Maybe the “craft” hasn’t disappeared, it’s just evolving into a new form.

[rproffitt]

Actually I'm seeing a morph to DISMAL MAGIC!

My son who taught A.I. classes is now working as a programmer and the next hire was nothing about programming. They call themselves a programmer but have taken no programming or related classes in any computer related field. All they do is ChatGPT and try to copy paste together "something."

[Reverend Jim]

Been that way for years. When I decided to leave my first programming job I told my boss that I should be present at the interviews. Being an engineer he assumed he knew enough to do it himself. He was wrong. Not only did the replacment not have a programming degree, he had only ever programmed in BASIC. The entire code base was in FORTRAN. It did not go well.

There are going to be lots of cases where unqalified programmers will be hired because "It's cheaper to hire someone to plug questions into ChatGPT." I'd like to see one of those execs on an operating table while the surgeon types, "How do I remove the gall bladder" into a ChatGPT session.

[Salem]

ChatGPT "programmer" == Shakespearian monkeys with an inference driven typewriter.

[DanielJack34]

Been that way for years. When I decided to leave my first programming job I told my boss that I should be present at the interviews. Being an engineer he assumed he knew enough to do it himself. He was wrong. Not only did the replacment not have a programming degree, he had only ever programmed in BASIC. The entire code base was in FORTRAN. It did not go well.

There are going to be lots of cases where unqalified programmers will be hired because "It's cheaper to hire someone to plug questions into ChatGPT." I'd like to see one of those execs on an operating table while the surgeon types, "How do I remove the gall bladder" into a ChatGPT session.

Yeah, that happens a lot. Hiring managers sometimes underestimate the technical side and it backfires. And relying on someone to just “ask ChatGPT” instead of actually understanding the work can create some pretty serious problems.

[Biiim]

I think your discussion actually hits on two key areas - one being the programming work itself and the HR department.

I think in any profession there is a tiered level of skill in it - the newer people usually take on the repetitive type roles to ease the workload off the more experienced who already know how things work and can think with how to solve more complex problems - but these new people should be learning and get able to do more eventually - even just by experience and without formal training.

In this profession as a programmer - actually understanding how to get computer program to do specific things - you know like foreach($arr as $k=>$v){ or for(i = 0;1< 5; i++){ - that's why it is a language and many programmers are multi-lingual and didn't know it. But aside from the language part of it, the job itself is very introverting - it makes you stop observing outside so programmers often end up with less social lives or trouble socially. I have programmed for over 20 years and it is very aggravating when someone interrupts you while you are in the middle of mentally following all the logic and how things flow and connect - then someone comes by and tells you it is really sunny outside and you just lost 5 minutes cause you got to get back into it again and remember where you were and what you were doing.

so when you look at programmers who have been doing it awhile, I am not saying this is all programmers it just has a large tendency to go this way, they are often painfully logical, very good at solving problems and often don't get the point of pointless social practices - you know like saying hello or catching up.

so then you get the HR department who are often larger-than-life talk to everyone kind of people - you know their social web is everything to them and they would go crazy if they couldn't talk to someone for 24 hours - they often have little knowledge of programming - or think they know it - but they have good skills at talking to people and knowing how people react to things - basically programmers can become very easy targets for them as they often lack in the social department or just don't care about it.

The creativity side of programming comes from the programmer himself solving the real world problems and doing the programming that solves it, and it gets destroyed with being told what to do and how to do it - I think it is the same for practically any job - the people who enjoy it most are very often the ones who are their own boss or have enough freedom at least to run their own area - and probably being successful should also be a part of it. If you want to get your creativity back stop just listening and doing all the orders and start questioning the orders and what problems they are trying to solve and solve their problems - likely the people giving you the orders at first will be surprised, upset, annoyed or shocked but often they end up liking you more and will want to work with you if they really want the company to succeed

Of course depending on your management and who you are working for this could be difficult to do - but the creativity comes from yourself and you need to get into the drivers seat of solving the problems and working out how to do it and not just doing what you are told. By problems i mean - rather that just accepting "we need a new sales interface" you need to find out why the person said we need a new sales interface - it might turn out to be one or a few people have said something and they just didn't know a facility existed or he deleted chrome and moved to Edge - he is trying to solve a problem too and you find out what that is - you are the expert on the programming so you have the best knowledge of how to fix it - but you need to know what is being fixed.

The best setup is to have a senior over the programmers and the programmers only listen to that senior and that senior takes all the requests/orders and works all this out and digs into things to get the real problems to solve - I would always do that for any of my juniors where I was working - I would tell them not follow what other people say and co-ordinate with me before doing anything on it - they got pretty good at it and other people in the company got used to the fact they had to tell me not the juniors and the juniors could get on with their work without distractions, you set up a team leader if you have a bunch of people in one area as it can get too much to check up on all of them, then you just check up on and help the the team leader and let him run the team members.

I think the AI apathy programmers have hit is the apparency that they are no longer needed but they just need to look into it a bit and get over that barrier of learning to use AI and keep on doing their jobs. I heard a lot of talk about programmers becoming redundant cause idiots can tell AI to program things for them and copy paste it, but the idiots are going to crash and burn soon or the idiot will become a programmer themself and cease to be an idiot. By idiot, in reality i just mean doesn't understand programming.

I like to look at it like a repeat of the industrial revolution with manufacturing making lots of people redundant, but it's not that programmers are no longer needed its just they can get work done 100x faster with this new tool.

[Reverend Jim]

when you look at programmers who have been doing it awhile, I am not saying this is all programmers it just has a large tendency to go this way, they are often painfully logical, very good at solving problems and often don't get the point of pointless social practices - you know like saying hello or catching up.

You pretty much nailed it for me. I enjoy saying hello and catching up, but once that's done I struggle unless the other person is a talker. In a work environment I got along well with others because the conversation was mostly goal oriented and farily structured.

I recall one meeting where the engineering group in IT called a meeting to discuss a new app they wanted to write for us. After fifteen minutes it was obvious they were fishing for an idea. I stood up and said, "This is a waste of my time. Call me back when you have something concrete." It didn't make me any friends, but my job wasn't to make friends, but apparently it was their job to waste my time.

Unless you work for yourself you will always have someone telling you how to do your job. You have to know when to ignore them. My boss did not like me doing things from the command line. Too bad. It's a lot easier to set up multiple network shares on twenty computers with a script than it is doing it from the GUI.

Software design and implementation is still a good mix of art and science, and even a smattering of psychology (for building user interfaces). Competent programmers are still very much required between the AI coder and the end user. Someone has to evaluate the code. You'd still want a meat surgeon to take out your gall bladder rather than an AI powered robot.

Point of interest, I saw just that on the news the other day. An AI robot removed the gall bladder of a live pig. The plan is, for remote communities, to have an AI robot perform routine surgeries under the guidance of a tele-surgeon.

[mandyDC]

That review part is probably going to become more important, not less. I was reading a recent Claude code review the other day, and one of the findings was that it could catch useful code-level issues, like some cross-file problems, but there were clear limits. It can review what’s in the repository, yet it isn’t actually running the application, so things like real database behavior, race conditions, or performance problems can still slip through. Broader architectural judgment still needs a human who understands where the product is going. So, AI may take more of the first-pass review work, but deciding whether the code is actually safe, maintainable, and sensible in context is a different problem.

[Dietrich_3]

I do remember the days when programming was very much a physical thing. You found students in the hallways with shoe boxes full of punch cards trying to get to the card readers. God forbid you got those cards out of order! Half the building contained racks of computer sections still using vacuum tubes, the other half had already switched to an ultra modern thing called a germanium transistor. A.I. was equivalent to a good chat in the coffee room! Good old days? I don't think so!

[Reverend Jim]

God forbid you got those cards out of order!

I still recall the blood-curdling wail of a classmate who dropped a five-hundred-card deck that was unsequenced. At least you never had that problem with paper tape.

Good times.

[Reverend Jim]

Creating a Möbius strip out of a paper tape for fun and profit

Ah yes. An infinite loop with only one side.

[Dietrich_3]

An exercise of the mind:
"What happens to a Möbius strip if you cut it right down its center?"
I had to physically do it, the result is so surprising!
Is their any application for it in data science?

[trcooke]

An exercise of the mind:
"What happens to a Möbius strip if you cut it right down its center?"

My mind gets fully stick thinking about this.... although intuitively I reckon you get two Möbius strips looped together.

[Dietrich_3]

To me A.I. is just a nice tool. Very helpful at times. I use it to stimulate my mind.
BTW that VW bus was the first car I ever owned in the early sixties! Same color, same model!