Programming as a profession is dead because of AI

Don't say that I'm not a professional and I just vibe coded a SMM panel and got my first sale. You just need to become a product owner/manager and not a developer.

Oh yeah, and then debugg. Devs still need to debugg.

Speaking of needing devs to debug.

I could use one

That means nothing.

I could code a blank white page and get a customer.

Come back when you have 1000 customers using your panel daily and see how you get on with vibin' it :-)
 
That means nothing.

I could code a blank white page and get a customer.

Come back when you have 1000 customers using your panel daily and see how you get on with vibin' it :-)
I didn't say that I said. I've made a panel and it's time for me to have a real developer clean it up. But you missed the part were I kinda asked for a tad bit of help. My panel won't have 1000 customers without some DEV love. Ive been in IM for ever. You don't get anywhere with free. Which is why I sold my motorcycle to fund this project.
 
AI is a tool, one that actually enhances the programming process rather than eliminates it. The core of AI itself is built on code, algorithms, and deep technical knowledge
 
I use LLMs to code, but I write and manage the code myself in sublime. I don't try to let them edit code in-place or manage the project, which is what 99% of people are doing. They can't do this. They're just pattern matching to solve problems.

They're actually better for larger companies with older codebases, because what you can do is fine-tune models on your past code/features/functionality to make adding new parts faster.

You absolutely cannot have "agents" writing code and putting devs out of work. It's laughable.

I'm the agent. The LLM, I just ask it to write a function and tell it what I want. It doesn't NEED to see the rest of my code. It doesn't even need to see the immediate file that the function will go in.

Let's assume I want to do some stuff with airtable.

If this is something I've not done before, I'll go to gpt 4.1 and ask it to give me a basic primer on working with airtable in python.

Now, I go install the lib, include it in my code and I know what to expect.

Next, I might say

"Write a generator function to return all fields in an airtable table, given a base, table, view and field=value to filter"

I'll check over what it gives, then I'll pop it in, test it, and I might then go back and say

"Add in exception handling."

I'll check that then add it in.

Then maybe I'll say "add in the ability to handle retries in the case of rate limiting and make sure to have a sleep timeout between retries of 30s"

Then I might ask it to write a unittest class with some tests, that I'll then place in tests/test_some_name.py. Again, I don't need it to create files, or do automatic stuff for me.

*I* am the orchestrator model.

I need to understand the code. No LLM agent is going to be able to replace a human like that.

If I don't know how to do tests, I go to 4.1 and ask "I'm creating a python app, and I want to learn about how I should do tests. Can you tell me about different testing methods, like which are essential, nice-to-have, and the correct way to set them up and structure my environment for deployment.

Vibe coding is definitely fun and has its place. It's great for non-devs to have fun with, but it has no place in a production environment.

Google uses a lot of AI to code, but they have so much code and infrastructure. They are definitely not using claude to vibe code stuff with cursor. ;-)

I've tested AI using other fields that I am knowledgeable of and I have to agree that treating it like an individual to do all the work for you is folly. It is good at pattern recognition but it has no idea what it's doing at all because it has zero understanding. The stuff it spits out is half right half wrong but sounds 100% right to the ignorant and lazy. It's easy to test this with a subject matter you're familiar with by drip feeding the AI with information to see how what it spits out aligns with what you know the result to be already.
 
vibe coders?
AI is helpful but real programmers specially skilled one is still in demand.
Just adapt and learn
 
I think it somewhat took some of the joy of programming. It made the easy tasks way too easy, and if I feel like if you don't do the easy tasks by yourself, you're not going to be able to handle the hard ones. While at the same time, the incentive to just put it to do the easy tasks is enormous.

But this is just my personal experience. I always learned going from easy to harder, and now it seems too boring to make the easy tasks myself (or time consuming), even though I know how to do it.
 
For now there is still a significant gap in quality when you compare something created by an experienced human and a newbie using AI to code

People with budget and a brain won't think twice and hire a human with programming knowledge using AI as an assistant, not as a replacement.
 
I just made a custom WP plugin during 30 minutes and charged my client $400 for a day of work.
AI is a power. You just need to make a very detailed description for your task.
 
Programming isn't dead! Or ever will be. It's just more fun, and quicker with AI. It still has its challenges as you say, that it gets things wrong from one moment to the next. Keep it on a short leash though, and its faster than search, copy and paste from stack overflow

I always now get it to write me logs for every function and variable call change, so I can seed back the logs back into it for quicker fixing. How I managed to work for a year+ without this method I have no idea. Lol even then I thought it was good, but when you add self diagnosing logs back into eh mix, its much easier.
This is a good idea
 
Yes, unfortunately it is not dead yet, but it will die soon. Even I, who know no programming, can code simple bots and professional websites thanks to AI. It seems that in 1 or 2 years, the need for software developers may decrease by half.
 
That’s like saying farming is dead because of modern farm equipments. It just made the job of farmers and those without niche farming knowledge easier, that’s all.
 
Yes, unfortunately it is not dead yet, but it will die soon. Even I, who know no programming, can code simple bots and professional websites thanks to AI. It seems that in 1 or 2 years, the need for software developers may decrease by half.

Coding websites and simple bots is a million miles away from developing enterprise grade software.

AI *cannot* do what I'm doing. I've tried. I would LOVE if it could. It just can't. Not even close.

I've tried many times. Every so often when there's supposed advances, I try again.

It's great at producing websites, because they're highly template based. It can add in simple features to those and make work-able apps for personal use, but as soon it reaches a certain size/complexity, it starts to struggle and introduce bugs. The only way to resolve those bugs is a human.

It's all non-devs saying that devs are cooked. That's what makes it funny.

You literally have no idea of the complexities involved in software development, and people that do know that there's just NO CONCEIVABLE WAY that transformer based LLMs are going to replace humans.

What they are great at is helping devs on-board to codebases FAST.

You can ask questions to a codebase and work out how it works really quickly.

Here's a slightly redacted version of the microservices architecture for my SaaS

1752958452097.png



Good fucking luck vibe coding that!

People saying devs are finished are like kids who have built a sandcastle, and think they can do the work of architects who design skyscrapers.

Coding little react apps with cursor or claude code is easy. The transformer is trained on public code, and there's a lot of public code for websites. This is a million miles from being able to create a full architected system with dozens of moving parts. It will just hallucinate a bunch of crap.

Without devs, the whole internet will literally die overnight.

LLMs are simply an aid. They can be used to help you build out functions/methods quicker than you can write them within your code, but only a dev can do this. It's like having a 3d printer print out parts for a car. Only a mechanic can put those parts together to build the car.

You admit. You know "no programming". You are in no position to to declare an entire industry dead. It's especially offensive that you say 'unfortunately'. People have spent decades perfecting their craft, and you want to just lazily end it all so you can code some slop while experiencing daily LLM-induced brain-rot.
 
Coding websites and simple bots is a million miles away from developing enterprise grade software.

AI *cannot* do what I'm doing. I've tried. I would LOVE if it could. It just can't. Not even close.

I've tried many times. Every so often when there's supposed advances, I try again.

It's great at producing websites, because they're highly template based. It can add in simple features to those and make work-able apps for personal use, but as soon it reaches a certain size/complexity, it starts to struggle and introduce bugs. The only way to resolve those bugs is a human.

It's all non-devs saying that devs are cooked. That's what makes it funny.

You literally have no idea of the complexities involved in software development, and people that do know that there's just NO CONCEIVABLE WAY that transformer based LLMs are going to replace humans.

What they are great at is helping devs on-board to codebases FAST.

You can ask questions to a codebase and work out how it works really quickly.

Here's a slightly redacted version of the microservices architecture for my SaaS

View attachment 460062


Good fucking luck vibe coding that!

People saying devs are finished are like kids who have built a sandcastle, and think they can do the work of architects who design skyscrapers.

Coding little react apps with cursor or claude code is easy. The transformer is trained on public code, and there's a lot of public code for websites. This is a million miles from being able to create a full architected system with dozens of moving parts. It will just hallucinate a bunch of crap.

Without devs, the whole internet will literally die overnight.

LLMs are simply an aid. They can be used to help you build out functions/methods quicker than you can write them within your code, but only a dev can do this. It's like having a 3d printer print out parts for a car. Only a mechanic can put those parts together to build the car.

You admit. You know "no programming". You are in no position to to declare an entire industry dead. It's especially offensive that you say 'unfortunately'. People have spent decades perfecting their craft, and you want to just lazily end it all so you can code some slop while experiencing daily LLM-induced brain-rot.
No, software engineering will be over and you will probably be unemployed. I don't care how angry you want to be, hahaha.
 
the AI hype reminds me of when people said self driving cars/trucks were just a few years away and would replace truck and bus drivers. that was around 10 years ago.

it reminds me of when internet first became big and people thought it would connect the world and make everyone nicer and kinder and more peaceful, the whole educated global village. it didn't.

when a new paradigm is unleashed, people are ignorant of its pitfalls and are overly optimistic. then as we learn more about this new paradigm we realize the limitations and reality sets in. this will be no different.
 
Computer programming is a skill that CANNOT be taught.

Years ago the government discovered two occupations that cannot be taught.
They must be discovered in people.

They are Air Traffic Control and Computer Programming.

I am not saying that coding cannot be taught. It can.

But being a software engineer entails MUCH more than mere coding.

It involves logic, planning, working with peers, working with upper management,
working with external clients, becoming an SME (subject matter expert), becoming
an application expert (like SAP or Oracle Financials), coding, version control, local
testing, group testing, quality control. You have to be able to take a functional
specification, convert it to a technical specification, convert that to a technical
design and then code it. Then test it. Then wait for QA (Quality Assurance) to
break it. Then fix it. Then QA it again until it passes. Then implement it at 1 AM
so you don't interfere with daily production. Then you "own" it so if it breaks or
needs updates because the users "forgot something" you work it again, over and
over and over. That is your career.

Most people cannot take that stress for long. Look at the personnel in any company
There are no old programmers

AI might assist a programmer by creating a framework to work from, if the prompt is good.

But anything produced will have to be reviewed in a DEEP step by step process to see if
it really does what it is supposed to.

AI CANNOT create new information. An LLM is a dumb. It does not actually reason.
It is a vast database of connections between words, That is how they are trained

If the prompt is detailed in what is being requested, not only the problem,
but how the code is to be laid out, commenting, strong data typing, division into modules,
division into functions, etc it can assist the programmer.

It cannot replace the programmer

Sincerely,

The Eye
 
Computer programming is a skill that CANNOT be taught.

Years ago the government discovered two occupations that cannot be taught.
They must be discovered in people.

They are Air Traffic Control and Computer Programming.

I am not saying that coding cannot be taught. It can.

But being a software engineer entails MUCH more than mere coding.

It involves logic, planning, working with peers, working with upper management,
working with external clients, becoming an SME (subject matter expert), becoming
an application expert (like SAP or Oracle Financials), coding, version control, local
testing, group testing, quality control. You have to be able to take a functional
specification, convert it to a technical specification, convert that to a technical
design and then code it. Then test it. Then wait for QA (Quality Assurance) to
break it. Then fix it. Then QA it again until it passes. Then implement it at 1 AM
so you don't interfere with daily production. Then you "own" it so if it breaks or
needs updates because the users "forgot something" you work it again, over and
over and over. That is your career.

Most people cannot take that stress for long. Look at the personnel in any company
There are no old programmers

AI might assist a programmer by creating a framework to work from, if the prompt is good.

But anything produced will have to be reviewed in a DEEP step by step process to see if
it really does what it is supposed to.

AI CANNOT create new information. An LLM is a dumb. It does not actually reason.
It is a vast database of connections between words, That is how they are trained

If the prompt is detailed in what is being requested, not only the problem,
but how the code is to be laid out, commenting, strong data typing, division into modules,
division into functions, etc it can assist the programmer.

It cannot replace the programmer

Sincerely,

The Eye

Completely spot on.

This is a man that understands both the world of software engineering AND LLMs.
 
Yes, unfortunately it is not dead yet, but it will die soon. Even I, who know no programming, can code simple bots and professional websites thanks to AI. It seems that in 1 or 2 years, the need for software developers may decrease by half.

No.

You just don't understand what real programming is.

I was an IBM Assembly Language programmer for 17 years.
I also was an SAP Technical Developer for 10 years

SAP is an Enterprise Resource Planning (ERP) system. An application

Wordpress is a simple internet application (CMS) Content Management System. An application.

Wordpress has 12 tables (not counting plugins you add)

SAP R/3 has 50,000 tables. FIFTY BLOODY THOUSAND. That is not counting
modular add on systems that can also be purchased (Customer Relationship Management CRM, etc

Can you see the difference between simple internet software and complex software SYSTEMS?

An LLM cannot reason or understand such a system because it cannot "understand" anything.

They are releasing AI at this time because as the number of boomers keep retiring and leaving
the workforce, there won't be enough people left to run "the system". So they have to do something.
The government has had AI for over 40 years. Top Secret. All that stuff.

We only get "trash technology" when it is no longer a threat to the system. ALL TECHNOLOGY IS WEAPONIZED

If you want a permanent high paying career, learn AI agents and go in to computer security. Get a security certification
of some kind, even a simple one like CISSP. Just to prove you have a brain. Even if you don't have a degree, you
will get a job. THEY will pay for the degree.

Read my previous posts

Sincerely,

The Eye
 
No.

You just don't understand what real programming is.

I was an IBM Assembly Language programmer for 17 years.
I also was an SAP Technical Developer for 10 years

SAP is an Enterprise Resource Planning (ERP) system. An application

Wordpress is a simple internet application (CMS) Content Management System. An application.

Wordpress has 12 tables (not counting plugins you add)

SAP R/3 has 50,000 tables. FIFTY BLOODY THOUSAND. That is not counting
modular add on systems that can also be purchased (Customer Relationship Management CRM, etc

Can you see the difference between simple internet software and complex software SYSTEMS?

An LLM cannot reason or understand such a system because it cannot "understand" anything.

They are releasing AI at this time because as the number of boomers keep retiring and leaving
the workforce, there won't be enough people left to run "the system". So they have to do something.
The government has had AI for over 40 years. Top Secret. All that stuff.

We only get "trash technology" when it is no longer a threat to the system. ALL TECHNOLOGY IS WEAPONIZED

If you want a permanent high paying career, learn AI agents and go in to computer security. Get a security certification
of some kind, even a simple one like CISSP. Just to prove you have a brain. Even if you don't have a degree, you
will get a job. THEY will pay for the degree.

Read my previous posts

Sincerely,

The Eye


Yeah, the problem is regular people are operating visually, so they see an LLM create what looks like a visually 'busy' app, and they associate what they see with architectural complexity.

Regular people judge by what they see rather than judging abstract representations of what a system really is.

You can see this with chatgpt. People just judge what they 'see'. They say AI is conscious because it's acting how a conscious being would act. They don't understand the abstract representation of what's going on in the model.

It's literally impossible for it to be conscious. Transformers can only ever estimate existing data. That data could be created by a conscious being, or an LLM, but the data itself doesn't inherently tell you what created it.

The only way that a transformer could properly simulate a human is if it was trained entire human history. Every movement, every action, every thought, every reaction since humans first evolved, and that would require computation and energy in the order of magnitude of a galaxy or perhaps more. Maybe even more energy than this entire universe has.

And if the transformer isn't 'conscious', then it can't be a developer, because to develop software requires conscious awareness and corrections based on being consciously aware. Not just blind predictive output based on input, no matter how many times the output is fed back in to simulate 'reasoning'. We don't function like that. We don't loop our output back in. We are aware of things and make decisions based on awareness. We don't have just 1 output based on what's in our awareness. We have the ability to 'feel' what's in our awareness and then direct the output in a number of ways.
 
Back
Top