I got made redundant after 25 years last year due to an 'advanced new system' which would save on wages. I got 2 years money finished on the Friday and started a similar job Monday.
i'm 54 now so if the company i'm working for pulls the same stunt in 3-5 years it will be brilliant for me.
Feel for some of my younger colleagues but for an old salt like me AI has been just the ticket!
But as tech types keep saying, there's no moat! Maybe Claude and ChatGPT are the best around for now
Thats not entirely true. The moat is the ability to throw money into a bottomless pit or come up with something really innovative to avoid that. For Claude is was fascinating listening to a sales pitch that it isnt just the model but how the models are integrated into the tools they are offering which is what counts. An attempt to start building moats compared to say github copilot/cursor.ai etc who try to be model agnostic.
Google and Facebook can subside their models for a long time.
I am not sure. Even they are playing weird and wonderful accounting tricks to move stuff off their main balance sheets. Their investors seem somewhat wary about the massive costs of the data centres.
Investors seem somewhat wary about how sustainable it is eg Oracles share price bouncing around due to their exposure to OpenAI.
Right now, "hey copilot, publish this review of new wheels" kinda works
It doesn't for all the reasons you suggested earlier. LLM (I wish people would stop referring to it as AI) don't know anything. They don't know what a wheel is, anything they write on any subject may be superficially readable, but they still hallucinate, if you ask them to review a passage or essay it'll mostly mark it's grammar, or review positively even if the essay is shit, it still mostly misses the point - because it doesn't know anything.
Outside of code writing, which at least according to this thread it does well enough, it's a best; a time saver, it can (sometimes) analyse a spreadsheet for you and produce graphics in a fraction of the time it would take a civilian like me to do them, but the work they produce still needs v careful checking as it just makes shit up, and it's very good at doing that. Yesterday the microsoft LLM told my ops manager that 3+9 is 11.Â
someone up thread called it an enthusiastic junior worker, I think it's like the dog in Up that has the device that translates its thoughts into English.Â
I suppose you've got things like spotting cancers in the plus column, but I don't think that's in the same category of techniques.ÂThey trained some pigeons to spot cancers with the same accuracy as experts a few years ago... So AI ain't that smart.
And my job will be ok for the foreseable future. As will a lot of my colleagues, we've got an entire team working on AI and LLM generally and another team looking at Gemini specifically, and neither of them are turning out anything that would be useable in the real world. Some of it just about reaches the level of garbage. A lot of it would be dangerous/illegal to sell.
It is useful for a lot of the programming tasks. Within quite limited boundaries.
I think it'll make my job (commercial pilot) slightly easier but won't replace me. Everything around aviation relies on deterministic provable programming, not an outcome which is different every time it runs. I'm sure you could teach the equivalent of Tesla's FSD software to take off and land, but you could never prove that it's safe.Â
Stuff like spotting hidden threats in weather radar returns is the sort of thing that takes a lifetime to learn, but a well-trained LLM could highlight them straight away.Â
IT Project Manager here. I can use AI for all the boring stuff and as a bit of a sounding board. But for the herding of cats and random requirement changes we get I think it would struggle.
Enterprise/ Solution architect here 👋Â
Similar thoughts to you
Same here. I quite often think if AI can do all of what I do, its welcome to it 😀
Was watching the news last night and they were interviewing 3 graduates who were struggling to find work. One a graphic designer another wanting to break into film and TV. Couldn't help think that they were truly screwed. Hopefully their horizons will broaden soon.
We think we've got problems as a bunch of 50 year olds. Fairly grateful I only really need another 4-5 years .
Statista has a chart on age bracketed unemployment rates, not yet at Financial Crisis levels but heading that way for 16-24 year olds. https://www.statista.com/statistics/974421/unemployment-rate-uk-by-age/ Â
I'm relieved to find STW being a haven of sensible assessment of this all. Reading the posts about the investment into AI, there is of course a huge incentive for those companies pumping $trns into AI to make companies believe that AI is the future and will replace all employees - and many business reporters are credulously reporting it. Even the Economist every week seems to have a minimum number of articles it must write on how AI is definitely really changing the world of work (nothing to do its Anthropic sponsorship deal).Â
But a couple of interesting things I read recently:Â
This https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it is a bit of a mixed bag. In the one company they observed, AI didn't make people redundant, but it did do enough of the menial, no QC required stuff that the employees could actually branch out into other things.Â
And on the Moltbook nonsense, this seems to sum it up pretty well:Â
One a graphic designer another wanting to break into film and TV. Couldn't help think that they were truly screwed. Hopefully their horizons will broaden soon.
I couldn't help feel that they needed to think about their career choices a bit more - especially the girl/lady who wanted to get into TV/films. My niece wants to do this also (has a great talent and has qualified very well) but the opportunities are slim - and even then it doesn't seem to be a good industry to be in. She's working making candles at the moment.
I really feel for all those looking for jobs, it's very demoralising getting knocked back time and again. So glad my 3 girls (23, 23 & 26) are all in good employment at the moment.
It's the nature of the "knock backs" that are the real demoralising factor for many young people looking to start their career. No human connection, reply or reason for rejection given. Young people spending hours, if not days, jumping through the hoops of application forms and online tests, only to hear nothing beyond an auto-reply.
it still mostly misses the point - because it doesn't know anything.
Some of the time that will come down to what you are asking it to do. Give vague requests and you get vague answers. Get more specific and it works better.
Which is possibly why it does better at software development within narrow parameters since you can have it tests itself with the "does this build?" which gets rid of quite a few of the more stupid answers. It does, which is going to be a problem longterm if management buys into it, struggle with anything new and poorly documented though.Â
Hark back to the time when we thought email would be time saving....
LLM (I wish people would stop referring to it as AI)
It is AI, it is Generative AI. Clearly very different from General AI but AI nonetheless.
It is really a pity that Keynes wasn't proved right especially as we are now almost in 2030;
In his 1930 essay "Economic Possibilities for our Grandchildren," economist John Maynard Keynes famously predicted that, due to rapid technological advancements and increased productivity, by 2030 people would work only 15 hours a week. He believed that once "absolute needs" were satisfied, societies would choose to prioritize leisure over endless consumption
Does look rather naive now with the way the world is run and by whom. Â
On a more general level and quicker than we think, AI will de-skill the vast majority of western workers. Not just in their jobs but across many many day to day tasks. This will have an affect on our brains which we can only guess at right now. I suspect we'll begin to think more like machines and less like individuals - group think will become endemic and we'll be lead by the nose as a species by the tech we have created by orders of magnitude more than we are now. People who say 'well someone has to program the AI' and 'we'll get the AI to do what we want' don't understand how it works and we are only in the foothills. Its basically going to coral the human condition down a particular pathway and whilst I suspect there may be a chance of steering this right now, the window of opportunity is pretty small and looking at those in charge it doesn't fill me with optimism as to what direction we'll be taking...
i draw things to solve simple problems on cigarette packets .... or laterally paint - then machine them.Â
from what ive seen AI isnt coming for those kinds of screw ups any time soon.Â
On a more general level and quicker than we think, AI will de-skill the vast majority of western workers. Not just in their jobs but across many many day to day tasks.
Won't the outsourced 'cheap' tech work sent to India etc., be the first to go?
Not for the foreseable.
Chartered Accountant in Education. New finance system is supposed to be AI equipped. Yeh right, we can't even get basic information out. Also AI doesn't have people skills so can't do my job. Should be fine until retirement
Won't the outsourced 'cheap' tech work sent to India etc., be the first to go?
Yes. The claims being made about AI resulting in deskilling by removing the entry level jobs has been in play for a long time now with offshoring. Of course this has had a visible impact over time.
Yep, expect 5 years absolute max in my current role in financial services. Already theres AI report writing and AI Financial Advice.
I think the bigger question isn't if you AI will make you redundant, it's the whole economy effect. There won't be customers or other jobs to move to if large numbers of the workforce haven't got money to spend. I'm working on a plan to protect myself from that outcome.
My wife works in insurance for the NFU.
Basically what they call personal lines - car, house insurance etc - is rapidly moving towards an AI only environment.
The bigger commercial stuff that she does is somewhat safer as it's a lot harder to send a bot onto a farm and do an assessment of their equipment, safety and buildings.Â
LLM (I wish people would stop referring to it as AI)
It is AI, it is Generative AI.
Nah. Saying it's something doesn't make it something - much like when "AI" spits out an output it doesn't make it true, useful or valuable.Â
It's a word jumbling machine.Â
It doesn't for all the reasons you suggested earlier. LLM (I wish people would stop referring to it as AI)
You're missing the point I think. Doesn't matter whether it's actually 'intelligent' or what we call it. What matters is that it can do the jobs of lots of skilled workers right now, and will soon be able to do the jobs of millions of others very soon. I despair for the younger generation. I have two daughters in uni (second and first year) and I'm increasingly concerned that there will be very few jobs outside of working in bars or supermarkets when they graduate. I expect to be supporting them for quite a while and I think that's going to be the reality for the forseeable future. For kids who don't have the luxury of parents with a bit of money that's going to be a catastrophe.
It's an important point though, because referring to it as "AI" is marketing spin... and an awful lot of effort is going into convincing people that these tools are "intelligent" rather than an effective way of reusing and repurposing existing work... and that's essential to stop people asking "why is the owner of the recycling machine getting paid, and not the creators of the prior work?"
Our lot appear to be doubling down, despite AI largely having been a failure to date. Worse yet, there is very much a sense of 'We have an AI shaped peg, so let's desperately try to find or create an AI shaped hole'.
Literally senior management asking us to spend our time trying to think up new ways we can use AI rather than getting on with our day job, as if they're willing to pivot the entire company to only producing things which, coincidentally, can be produced by AI. I doubt it will fly in our particular industry which basically relies upon critical thinking, accuracy and detail 🙄Â
Daughter is just doing her Masters in Digital animation - potentially a worrying area, but I don't think AI is likely to be capable of the 'personal creativity' bit.Â
Mechanical maintenance on an oil rig. Not a chance mine will get done by AI although the amount of bloody paperwork we need to do now it would be welcome to do that side of things and let me actually fix things.
It's not replacing me as such - but it is making my (design & creative based) job very very dull and unfulfilling ever since my boss has decided that it will increase output and we all have to use it as much as possible.Â
Things I used to enjoy like layout design, copywriting etc have largely turned into me shouting at a robot and then making tweaks on the dogs**t that it spits out; rinse and repeat. I know it saves time but it makes me feel worthless and stupid, and somewhat annoyed.Â
It's getting better at quite an alarming rate though, so I would imagine in a few years, if the bubble doesn't burst, then it would outright replace jobs like mine.Â
In our customer service team, new staff are also being told to heavily use LLM answers & info, and therefore gain absolutely no technical or product knowledge compared to if you actually learned about the thing you're trying to help someone with.Â
Luckily I have just got a new job doing something totally different!
I'm increasingly concerned that there will be very few jobs outside of working in bars or supermarkets when they graduate.Â
Have you tried getting a tradesperson recently?
but I don't think AI is likely to be capable of the 'personal creativity' bit.Â
The question is how willing is management to be paying for that personal creativity bit? Whilst in some cases they will in others they will decide its good enough.
Although again this does go back to how sustainable the growth is and what the price model is going to end up being for an AI tool.
For a time there was the joke about AI standing for "Actually Indian" due to several companies claiming advanced AI tools in reality using offshore teams. Once the token cost hits a certain level humans are going to start to appeal again. Whilst some companies seem to have improved model efficiency a lot of it does seem based around just throwing hardware at the problem and then wrapping it in various guardrails/prompt rules to make it appear better.
As the great Charlie Mullins said recently "Learn a trade to avoid AI jobs bloodbath". Probably the first and only ever time I would agree with him seeing as he is such a massive ****.
Have you tried getting a tradesperson recently?
Who is going to hire those tradespersons though when everyone but them, barstaff and a few other professions have been undermined?
Time to bodge it instead.
I'm increasingly concerned that there will be very few jobs outside of working in bars or supermarkets when they graduate.Â
Have you tried getting a tradesperson recently?
That'll change when a large percentage of the workforce have less income/unemployed
I use LLMs a lot in both work and private scenarios. With familiarity you become very aware that it is isn't thinking. It is reflecting.
I know this is not where AI development ends, but the idea that you can get a detailed and pertinent answer out of AI when you don't provide clear instruction is fallacy. All the futurists predict a watershed where delegation to AI starts to be become viable - at that point you will need to meta-manage the AI to ensure its intent is aligned with yours or... surrender. That's the AI apocalypse.
Right now, I can be more productive because of my interactions with LLMs. Compared to human peers and subordinates, I can iterate an idea on a topic quicker and with a better access to supporting information.
But, the limitations are worth keeping in mind. The responses can provide lots of false information with astounding confidence. There can be errors through omission. In essence you need to know enough about the subject area to spot the LLM blind spots.
So right now I'm feeling confident. My personal productivity is up and my value as an employee is high enough to see me to the end of my career. Â
maybe someone in construction has a reason for that?
Seems to be, in the UK, a mix of bad press due to some of the post WWII variants, planning issues and mortgage issues.
There is also the problem that they come with large upfront costs and whilst there have been a few attempts think all have failed now due to not having enough orders.
Its the sort of thing which a coherent government strategy could address but sadly we just have them shouting about newts blocking building and trusting the big housing companies to completely change their strategy and undermine their own profitability if we relax planning rules.
Who is going to hire those tradespersons though when everyone but them, barstaff and a few other professions have been undermined?
Seemed to be a fair few posters saying it wouldn't be making them redundant up above. Interesting how some roles, eg IT coding, which have only been around as mainstream jobs, say thirty years will be gone.
Who is going to hire those tradespersons though when everyone but them, barstaff and a few other professions have been undermined?
Seemed to be a fair few posters saying it wouldn't be making them redundant up above. Interesting how some roles, eg IT coding, which have only been around as mainstream jobs, say thirty years will be gone.
Any job that follows a model or rules will reduce in numbers. Law, Architects, Massive efficiency gains for Doctors, Call Centres, Purchasing etc...
I've used AI a few times at work, but don't really trust it & can't rely on it.
Some of the managers at work are really trying to push us into using it for day to day stuff. I didn't want to come across as a completely dismissive luddite so have tried to embrace it.
The first time I used it, I asked it to tabulate the properties of two refractory metals so I could compare them easily.
I figured this would be quicker than me finding datasheets & putting the info into a table myself.
It got the very first output row in the table wrong (melting point) even though when I looked at the source material it referenced it was correct. So, it had somehow reversed the numbers in the output which meant the rest of it could not be trusted & I had to check it myself.
The second time I used it was for some broad outline concepting work. It created ideas that sounded feasible but I couldn't quite get my head around what they might look like. No worries, it asked if I would like a diagram to accompany the ideas to help me visualise it. Sure, sounds great. Yes please.....
The output was complete nonsense & while some bits made sense; all the difficult bits which my questions were really targeted towards were complete visual gobbledy gook. Like it had just cut and pasted images from all over the place into something that it thought made sense.
Quite possibly there will be a reduction in numbers in certain disciplines within work areas, such as junior contract and commercial law, but in others areas not .
Maybe I'm just an optimist. But I for one salute our new AI overlords.
From this year's Chinese Spring Festival TV extravaganza:
Coming soon to a street protest near you:
