‘You cannot stop this from happening:’ The harsh reality of AI and the job market – “I’m really convinced that anybody whose job is done on a computer all day is over. It’s just a matter of time,” one engineer told Michelle Del Rey
https://www.independent.co.uk/news/world/americas/ai-job-layoffs-tech-unemployment-b2769796.html
46 Comments
From the article
The often-talked [threat of artificial intelligence on jobs](https://www.independent.co.uk/news/world/americas/college-graduates-2025-job-outlook-ai-b2758693.html) suddenly became very real and shocking to Jane, who asked to use a pseudonym for privacy reasons, when her human resources role [became automated](https://www.independent.co.uk/news/business/klarna-ceo-sebastian-siemiatkowski-ai-job-cuts-hiring-b2755580.html) and she was laid off in January.
She’d spent two years at her company managing benefits and was on track for a promotion. She’d noticed her boss [building out AI infrastructure,](https://www.independent.co.uk/tech/meta-ai-artificial-intelligence-privacy-facebook-b2769787.html) but didn’t think her position, which paid roughly $70,000 a year, would be affected.
“I thought that because I had put in so much time and been so good on the higher-level stuff, he would invest in me,” the 45-year-old Bay Area resident told *The Independent* about her former employer. “Then, as soon as he had a way to automate it away, he did that. He just let go of me.”
I worry most about young people.
They get degrees with the expectation of getting an office job. Companies are already freezing new hiring.
Young people not working will affect everyone, young AND old.
Automation can be resisted with job-creation policies if we choose. But is that the outcome we want?
We should be talking seriously about UBI. The purpose of the economy was never just to keep people busy; the economy is supposed to help us prosper.
There’s plenty of living for us to do in the absence of jobs.
Reality is not that simple. LLM do not work reliably alone. They hallucinate, make mistakes and fail.
Automation destroyed a large number of jobs, and new jobs were created. Same will happen here with LLM.
Jobs that effectively did nothing meaningful and were the result of population explosion, were never going to last. It was a temporary “fix” to delay the problems.
Stop this bullshit with trades. Trades suck. Most of them still don’t pay enough. I know your uncles friends son is making 1 mil a year in the trades but guess what? that’s not everyone. not even close. There’s also people in white collar work making multi millions and billions of dollars but that doesn’t mean everyone in white collar is making bank. Not to mention of you send an entire generation into the trades the pay is gonna plummet along with it.
AI is gonna bite businesses in the ass when they wake up one day and realize there are no experienced people to cover what AI cannot because they stopped letting people gain experience years ago.
I love how many people don’t believe this. They think is fear mongering or as I’ve been told “you need to adopt to the change”. I work in tech, we all clearly see how our jobs are being replaced.
It’s a business decision, Completely personal.
This still isn’t a great take. Yes, AI can code. Yes, it can automate some simple and repetitive tasks. And yes, some job loss will occur. But the scale of disruption being pushed in so many of these articles is significantly overblown.
Take Microsoft, for instance: they’ve stated that up to 40% of new code committed by developers using GitHub Copilot is AI-suggested. But that doesn’t mean Copilot is autonomously writing Windows or mission-critical code. These are suggestions accepted by human developers, and a lot of it still requires cleanup due to redundancy, inefficiency, or even incorrect logic. It’s helpful, but far from reliable.
There’s also growing evidence that AI tools frequently “hallucinate”—they generate incorrect or nonsensical output with full confidence. This has serious implications: in mental health tests, for example, some AI-powered systems have given harmful advice to users, like suggesting they stop their medication—something no responsible clinician would say.
Executives will absolutely use AI as a justification to cut headcount—we’ve already seen it. But many roles will change more than disappear. Research from MIT and Stanford consistently shows task automation, not full job replacement. In most industries, AI is better at augmenting work than replacing the worker entirely.
Bottom line: These models still require close human oversight, especially from domain experts. Trusting them blindly, whether in software development or high-stakes environments like healthcare, is not just naive—it’s dangerous.
The real problem is that we are culturally unprepared to pay people for doing less or nothing. We have to take the enormous profit of automation and be willing to share. And then have people seek value in their lives beyond work.
Back in the day, a “computer” was a person that was paid to sit at a desk and do calculations. Companies would hire rooms full of these people to do maths for them. When digital devices were invented, these jobs were no longer needed and the world moved on.
People be freaking out about AI, but this happens like every decade at this point. We’re going to be fine y’all
No shit. On a long enough timeline its a certainty. Now…what that timeline looks like is what were all discussing and disagreeing on.
There are engineers who solve engineering problems of the kind for which a single human is sufficient. The kind who stereotypically have poor social skills and do their jobs in isolation from others.
And there are engineers who solve problems too complex to ask of a single human, and part of whose job is to work with groups of people.
Guess which kind this one is.
(It’s horseshit. AI sucks at dealing with complexity or comprehending the implications of its decisions in a multitude of engineering fields where training data on real consequences of decisions simply can’t be gathered. It is categorically unfit to make decisions where safety stakes are high, from aerospace to medicine to education. AI has applications in numerous tool chains including in these fields, but applying it will continue requiring humans, many, **many** decisions will not be handed over to AI, and this kind of drivel is nothing more than sensationalist knee-jerking).
In short if your job is a series of if then statements which follow a very precise process and only require escalations for fixed criteria yea.. you might want to look into something else
Creativity and ingenuity will reign supreme until a new set of roles are created.
They confuse the communication/interface with the work and effort. A wise man can point to the moon and a foolish one will stare at the finger.
All these engineers think they’ve cracked it… and it’s their jobs that’ll be the first to go. 😆
My phone can still not reliably transcribe when I talk into it. Someday maybe. Not anytime soon.
My 30-year marketing career in design, interactive, animation, and 3D flushed down the toilet in a matter of months. This was *already* 2 years ago.
If that’s true, who’s going to check their work? AI is inconsistent at best, wrong at worst.
Why does it feel like this is the only genre of article posted here now?
I was a lot more impressed with AI a few years ago, when ChatGPT was new. It hasn’t advanced nearly as quickly as I thought it would then, and anyone who actually replaces employees with it is in for a rude awakening. I think it’s a lot more likely that AI ends up being used as a tool to make employees more efficient and productive, which actually ends up creating a demand for more employees, as we saw when computers entered the office space.
All of these companies downsizing with AI aren’t going to have anyone to buy their products or services. We’re either going to need to restrict firing people to be replaced with AI, or have some kind of UBI. Maybe there are other solutions. But if 90% of us are unemployed, there won’t *be* any companies who can lay people off.
Have been using a paid version of chat gpt the last few weeks and I can absolutely see this taking the jobs of a LOT of people in the next several years. It’s still needs a bit of help with nuance, certain emotions, and a few other key areas but it’s well beyond a baseline. It’s incredibly scary how good it is I can only imagine others catching up and where this will be by 2030
I always say what happens when we’re all out of work. They have to give some type of money. But they won’t.
Bold words from an engineer whose work is done on a computer all day
AI requires CONSTANT human data to feed on. There is no way to fully replace people but they will 100% try to avoid dealing with the class war that’s brewing. The ruling class hates humanity, they even hate their own humanity. Look at billionaires like Elon Musk and how they talk about people and themselves. They are so ashamed to be human that they’ve decided to make everything sterile and lifeless to avoid feeling anything at all. Being a billionaire is a mental illness.
Yeah, this is mostly bullshit and a good narrative to push when you don’t want to spend the money to hire new people really AI is not gonna take over anybody’s job for a good 10 to 20 more years as it now is only in its infancy the one thing that will be a problem is if you do not learn how to use AI efficiently with your current job. You will be out of job in five yearsas long as things keep progressing in that direction.
AI has repeatedly shown its only as good as its programming. Codes and program fail all the time. This makes AI stupid and without knowledgeable people to fix it, it makes everyone more dumb.
Anyone who dares question AI’s ability to take everyones job is seen as an AI skeptic like people not working is the endgame for AI. 🤦🏾♂️
I mean if a job can be automated is it really worth a human doing anyway? It would seem like a waste of your life doing something that a computer can do cheaper.
Just in case any of you thought UBI was going to come save the day, it won’t. David Sacks, Trumps AI czar, called UBI a “fantasy” that “is never going to happen.” Seems pretty unequivocal to me.
We can’t get basic healthcare or community college. Do you think the owners of this tech will voluntarily pay money to useless eaters? Forget about retraining programs, too.
Where are we going here? Seems like AI is going to cause way more problems than it supposedly solves.
This is stupid. People will be doing different work with computers. There will be a transition period and retraining required to get there, but it’s not the end of jobs.
It’s always been a matter of time, the question is how much time?
The LLMs we currently have are FAR from being able to replace skilled jobs entirely. At their best right now, they can semi-reliably automate some tasks (with close supervision from actual humans who know what they’re doing).
I think before we make AI that is remotely capable of replacing skilled workers, it’s an order of decades, not 1-2 years like AI companies say every year.
And they mock me for being a bartender….it’s like being a hairdresser, we cannot be replaced by robots.
And the internet would replace all of us! Meanwhile it’s created no value!
What about people working on a computer all day developing AI?
I’m hoping it turns out to be a redundant fad much like the NFT hype a few years ago. Probably not, but I’m still hoping it dies.
Most companies laying off office workers in Western countries aren’t replacing them with AI, they’re being replaced by cheap labor in India.
I see so many articles and comments on this sub about what happens to workers because of AI, but nothing about what will happen to businesses.
If every business uses the best ai for their industry, won’t they all be using the same tool? And if that happens, then isn’t competition dead? If this technology is really so powerful that it can replace our coders, writers, designers, and what-have-you, what will stop the owners of the AI from taking over all business? Aren’t business leaders essentially giving their entire ip and model over to third parties by handing the reigns over to AI? It’s not like AI is unowned and a free agent or even a tool – it’s a product being sold by a business.
Tl;dr – why aren’t business leaders worried about their companies going under once the owners of AI decide they’d rather steer the ship?
Whoever made that quote hasn’t programed on unreal engine
Yes we can, if government can intervene and we can hang on to our copyrights
The western world has the most to lose from Ai, the east is about to completely brain drain back all of its literal knowledge and culture — somewhat poetic considering the colonialist past
Profit-driven businesses have always enshitified and will always continue to do so. Their owners don’t want employees, they don’t want you, and they never did. They don’t want to provide a good service and they never did. Profit is the only thing that matters. They will replace all employees with automation even if it means their services degrade (which they will, without humans operating them), as long as they keep profiting and growing. Soft skills won’t save you.
Jobs shouldn’t be part of a market. That’s why we’re in this horrible situation.
“You cannot stop this from happening,” says person in a society where this is only happening because shareholders are prioritized over every other form of human life.
We *can* stop this from happening. We’re just not going to.
I think some maybe most people have the wrong idea as to who and when for AI and the workplace…
Our current AI and most likely near-term AI. Wont replace those making something new. Programming for example wont be replaced by AI. The work towards new will be done by people. The work that’s been rehashed a million times will be done by AI tools to make repetitive work easier and faster. This WILL lower the number a given company needs by a non-insignificant number of people still overtly affecting those in the programming field….
The when is more a question of how long it takes for the tooling to perfect a given repetitive task or answer to a question. It cant just give an answer it has to be correct for it to be profitable. In the vein of call center workers. This means the front line no longer needs to be someone suffering in answer to every single call. But every single call could be used for AI to learn how to take those front-line questions. Machine learning is rapid growth with bigger models and though it doesn’t work well this moment. It can work well in the very near future as growth from data points is exponential. Edge cases and more complex issues will still need to be triaged by a real person at least for now. But that does mean companies will need fewer folks to do this work over the next few years. Overtly affecting those in any field that directly interacts with consistent and repetitive tasks.
It will however cripple our society if we don’t move away from the old social contract. If we don’t start setting people up as valuable not because of what they can produce but because they are valuable as people. We need leadership who are thinking about the life of the average person tomorrow. Having discussions about the very real realities of what AI will mean in the very near future and preparing for it adequately as folks leave the work force.
LLMs are a tool that make people a bit more productive. Kind of like when google came out.
They are not going to replace very many people, and many of the implementations where they do replace people will be rolled back once they start fucking up, which they do, frequently.
I think 100% of the people making these predictions really do not understand the technology, the industry, nor how LLMs actually work in fact.
I think these folks do not understand LLM is nothing at all like AGI, and we are decades away from AGI.
I’m on a computer all day building autonomous systems. I’ve been doing that since 1999. I’m fine, you should stop that group think stuff. Some of us planned for this
Companies are already pivoting away from AI. At least the smart ones.
They’re finally beginning to realize, at least the smartest of them, that AI is not what it was promised. It’s a pattern recognition engine and it will more than happily repeat the pattern of a consistent error or hallucinate a false solution that sounds like it best matches the pattern of consistent error.
AI is great for making sure that it only drills through the parts and doesn’t drill through the bracket holding the parts.
AI has no motive and therefore no creativity.
And there’s enough bad code out there with proposed solutions based on that bad code AS to make one laugh.
The weirdest suggestions I get from the AI variable completion system that they were pushing at work would produce laughable results. Customer would be laughing but anybody examining the code would be.
And it looks like the AI people are now potentially looking for a different AI solution for code completion etc.
Fundamental problem is that large language learning models are not intelligent, they simply sound intelligent on TV.
If you ask them a very specific question that can be answered by a search in pattern space then they can find answers nobody thought of because people can’t hold that big of a pattern space in their head.
But you need just as many people to ask the very specific questions as you already need to do it yourself. And to some extent those people have to be even smarter because they have to be able to vet these suggested solutions.
Over on some of the Linux groups people talk about asking AI how to do something. Trying it and getting a terrible system damaging result. And then telling AI that solution you offer me doesn’t work what else should I try. And then it looks for the pattern match of all the people who know that the first popular solution is crap he gives you the solution should have given you in the first place.
So like if you ever get a solution from AI, you should say and when that solution doesn’t work what should I try a couple times to see if it comes up with better and better solutions. But be on the lookout for it to come up with worse and worse solutions just as easily.
Nah, tech support will still exist, to untangle whatever mess an AI has implemented after a badly phrased request or something.