Engineer says Claude Code has made his job "soul-sucking" as workers spend 12-hour days pressing enter

midian182

Posts: 12,048   +184
Staff member
A hot potato: One of the many things we were promised that AI would enable was making coders super productive. But one software engineer's viral post claims it is reducing employees to little more than unskilled drones who work "12 to 13 hours a day just to press enter."

The engineer, who posted anonymously on X under the username voxium for fear of reprisal from his employer, called his new role at an undisclosed company "soul-sucking."

There are doubtlessly quite a few people who would describe their jobs the same way, but the source of voxium's misery is Claude Code.

He says that Anthropic's AI is now generating everything from product specifications and tests to tickets and reports.

"Nobody knows anything here," voxium wrote on Sunday. "People are working 12 to 13 hours a day just to press enter."

The engineer added that employees are now pressured to ship products as fast as possible, leaving them with little time to review Claude's output or understand the code being created.

Voxium said he wouldn't mind the situation if workers were given more time to check the code.

"Everyone, literally everyone, from an L1 to an L7 engineer here is doing the same thing. Talk to Claude," he added. "There is no sense of victory. Nobody is resolving bugs. In reality, nobody is thinking anymore."

Some people might look at this and wonder what the problem is – who wouldn't want to be paid to do very little? But most people struggle when their work has no meaning. Voxium told Business Insider that if engineers lose the opportunity to understand the code and solve difficult problems, "there is nothing left ... Not even our soul. There will be no purpose."

He added that the real problem is corporate leaders' obsession with pushing out more features. He said management often judges success by sprint cycles, pull requests, and the sheer number of features shipped, even when those additions do little to improve the product for users.

Several other coders who replied to the message said the situation mirrored what was happening at their own companies. In addition to facing the same loss of purpose, some fear it will lead to job cuts as Claude takes over more tasks in the coding workflow.

The post received plenty of replies from high-profile figures, including Elon Musk, who added his favorite "Yikes" exclamation.

Chamath Palihapitiya, co-host of the All-In Podcast, warned that AI was turning a generation of technical experts into people "pushing a button in front of casino slot machines."

Back in 2025, a study showed that AI coding assistants were actually slowing down experienced developers. Since then, we've heard how AI coding is evolving into "loop engineering" – a system that prompts itself.

As a reminder that AI coders can get things very wrong, one agent in April took only nine seconds to wipe a startup's production database and its backups with a single API call to its cloud provider.

Permalink to story:

 
It can be hard to understand one's own code from a long time ago, so it is improbable that programmers are going to understand generated code with little time to reflect on it. The mental faculties of programming will, I think, slowly atrophy without use.
 
99.9% would happily pocket a 6-figure salary + assorted perks to press Enter, so the misery of a some spoiled attention seeking individual hardly deserves any attention.

That is ... if he/she were a real person.
This is an obvious troll post, all these unverifiable claims sound ridiculously absurd to anyone who ever developed software. Churning out code has never been the most important part of software development, and the fact a significant portion of it is being automated can't be a reason for complaints.
 
Before you ridicule him as "moping elitist high salaried tech bro", know that tech workers right now are the canary in the coal mine for AI.

AI models were trained first on code, and programmers were the first and most enthusiastic users of AI automation. But what the OP is describing is spreading to a lot of other white collar professions. You don't need as many "knowledge workers" when AI has all the knowledge in the world.

And eventually it will spread to any job that only requires minimal thinking and simply pressing a few buttons. I was at home depot just yesterday and there were plenty of staff but whenever you ask them for something they just pull out their phone and look it up on home depot website. I can do that myself! The registers are already mostly automated.

Regarding software, I'm not sure that what we were doing before was that ideal either. There were plenty of people who didn't understand the code, or read code reviews, and bugs everywhere, and execs who just wanted to ship features and not bother with technical debt. It's the same old story, just massively turbocharged.
 
Don't fret tech-broham, you aren't just "pushing enter"; you are serving as a "meatbot responsibility anchor (tm)" for Claude's work. You'll discover your real "soul's purpose" when that code you just "pushed enter" on humps the pooch.
 
Mindlessly pressing enter and trusting AI will lead to disaster.

I use Claude a LOT, but stay within the $200/mo plan constraints because I have to review what it's doing and make corrections. The human is the bottleneck and the only thing that makes AI's "leverage" useful and not just slop.

Still, I can offload enough drudgery work to AI that I'm about twice as productive even with supervising AI. That only works if you are an independent contractor (or similar) to take with wins home as time or money rather than your boss just upping what they want.
 
This is not wrong. I am forced to use AI for my work, but I do my own edits. If you speak with IT technicians here, they are either Book smart or AI smart, with no real experience. I wanted to order SO-DIMM single-rank DDR3 memory; the technician told me you only get 1 type of ddr 3... This is a simple thing you already learned in the 2000s. They also do not work on any old hardware, which is a bonus for me, as I specialize mostly in old hardware..

AI made people lazy; in school, you do not learn how to use your hands; you learn how to use AI. For kids to pass here, you just need 30% to pass. You hire people who have no experience. They get a massive salary just to use AI each day. For me, that is hell boring thus I do my own edits manually. Could not care less what my manager says.
 
The problem here doesn't really sound like Claude. It sounds like management discovered AI lets them demand 5x the output and immediately converted every productivity gain into more deadlines.
 
This article is completely out of touch with reality. And if you don't think so, then you clearly don't understand how to build software with AI.
 
As a Dev, I use Claude all the time and my experience has been quite different, I set it going about a task while I work on my jira cases, It has allowed me to be rather productive, Yesterday for example I had Claude working on a new project while I was online with another dev helping her with something else. So I still got the job done I would have been doing while helping on another project.

It has reduced my cognitive load and taken over some of the more tedious tasks, It takes time to set it up so that it works how you want but it's worth it.

So I am more productive and I am still writing my own code, It's working to the terms I laid out so I can keep enjoying my job. Life is what you make it I guess.
 
Why are they working 12 hours per day though? The parts I agree with is when everything is forced to be done by AI without human thinking - when management wants everything to happen at a magical pace because AI can do everything faster.
 
This article is completely out of touch with reality. And if you don't think so, then you clearly don't understand how to build software with AI.
I am afraid you are mistaken. Some companies do not allow developers to use AI in a productive/common sense way - they force staff to use it for EVERYTHING. Developer feedback does not matter to them.

- Jira ticket comments are flooded with inaccurate information from AI tools.
- Confluence documentation is contaminated with more inaccurate AI vommit, which is also not being maintained/updated as things change. If you search for something in Confluence, you cannot trust it or it's full of AI vommit output that misses the things you'd expect it to have.
- No one can trust anything they read as being correct.
If you are caught not using AI, you may be in trouble... that's what I deal with in a tech company.

Even products & front end user interfaces are designed by AI - I took just one screenshot and could identify 5+ flaws in the entire concept, not just the interface..But the people responsible for applying corrections cannot get to that because they are expected to output insane amounts of work for 1 or 2 people by using AI.
 
I'm looking forward to the time when Claude et al can more easily permanently erase, without any chance of recovery whatsoever, whole corporate databases (including back-ups) which have taken decades to accumulate. Sweet.
 
This is not wrong. I am forced to use AI for my work, but I do my own edits. If you speak with IT technicians here, they are either Book smart or AI smart, with no real experience. I wanted to order SO-DIMM single-rank DDR3 memory; the technician told me you only get 1 type of ddr 3... This is a simple thing you already learned in the 2000s. They also do not work on any old hardware, which is a bonus for me, as I specialize mostly in old hardware..

AI made people lazy; in school, you do not learn how to use your hands; you learn how to use AI. For kids to pass here, you just need 30% to pass. You hire people who have no experience. They get a massive salary just to use AI each day. For me, that is hell boring thus I do my own edits manually. Could not care less what my manager says.

On the other hand, people who were afraid to do things can now learn them faster and gain new skills.
I am talking about physical tasks, though.
But that is the next thing, right? Robots with AI that can do everything humans can and do it faster, more accurately.
 
If this is true, I imagine Claude is just training itself through machine learning to understand when to press "enter", and then tell the company they can let these employees go because it figured out an algorithm that can omit the need for them to press "enter". lol
 
I code for a game I play on the side, and boy… I am in the same boat. I don’t get the enjoyment out of my hobby anymore. If I weren’t to use AI, I wouldn’t stay competitive in this day and age.
 
I'm hearing that architecture is the same thing, CAD and kit projects make the job soulless and boring.

I don't know where you heard that - because it sounds like you just made it up, or don't know anyone in the industry. CAD and CGI are foundational to the arch/design field, and there's no way any "AI" could ever replace me or my structural engineers. It's simply not possible, because they have no autonomy, no intelligence, and can't look at or analyze structures to begin with.

I have met a handful of "designers" who attempt to use "AI" to help out, and it gets everything wrong, every single time. It cannot even count, much less calculate materials properly or measure actual, real structures or space. It's hilarious.
 
The powers that be ($$$) are shoving AI on all of us. They need to get a ROI for all the money. I foresee a time when AI will diagnose what is ailing without the help of a Doctor or APN. What happens to all the people who lose their jobs across all salary ranges. Ultimately it will lead to violence and revolution.
 
Back