Hands-on jobs to grow fastest, because AI can't touch them Think tank and advocacy org the World Economic Forum has predicted strong growth in jobs that AI can’t replace, plus big demand for skills to automate those that can. The org (WEF) this week dropped its annual Future of Jobs report [PDF], which is based on a survey of over 1,000 employers who collectively represent 14 million- …
They could replace some of HR, though that might just be simpler by an if-else-flow, though less easily deniable...
"if $(worker) request leave {
if $(worker.status) == troublesome worker.fire()
else $(worker.status) = trouble
fi
fi
or something like that. Having "AI" in it would make it deniable, ie. telling the tribunal "honest, guv, we wasn't ta know the machine would do that!"
In Canada, corporations are required to abide by any agreements their AI enters into. Witness one Air Canada customer who got the price the AI quoted her, not the proper price. That's enough to make corporations here think twice about turning things over to AI that haven't been proven to work properly, because it can cost them money and liability under Canadian law.
This post has been deleted by its author
> Hallucinating is the daft Silicon Valley term. Us users of AI prefer the term 'bullshitting'.
Bullshit has real value: to plants, to dung-beetles, and dried as fuel (Lisu, تپی , Кизяк, Ghunte, Gomaya, Muttal, Кизяк, etc). Cow Poop At HAAS
Do not tarnish bullshit's image with useless hallucinations.
The obvious roles to replace are the over-paid, under-talented twats of the boardroom. Think about it, you (almost) all know what board reports are like: Synopsis, options and recommendation, then a meaty report directors are TOO FUCKING LAZY* to actually read, followed by a modest amount of back up materials.
We don't need pompous twerps in expensive suits to make the choice, the board papers are perfect for feeding into an AI, and asking "what is the optimal decision?"
* For those of you who do or have prepared board papers, can you think of any occasion where there's been even a vague smell that any board member bothered either with a pre-read, or even to look at the papers at all before they entered the room?
This post has been deleted by its author
These are already sectors where there is a chronic shortage of labour, partly because of an ageing population and partly because public services are increasingly underfunded. I'm not sure how creatng more unfilled vacancies is going to help.
Don't worry people (or perhaps you should worry)
Elon the Bountiful will be onto it when Trumpo gets fed up with him and [drumroll please] gives him a pink slip and says, 'You're Fired'.
He will make robots that can do everything we can do... yes including that. Or that is what he said before he became all MAGA.
The US will need them if the OJ deports even 5% of the immigrant workers otherwise the MAGA cult faithful will not like it when the price of eggs doubles.
The price of eggs may not double, industrial sized egg farms already have a lot of automation. But strawberries, broccoli, lettuce etc. could not only double but would also be in very short supply as crops rot in the ground with no one to pick them
>But strawberries, broccoli, lettuce etc. could not only double but would also be in very short supply
So what ?
Only democrats eat vegetables. So long as a big-mac and a mega mac-slurpy-gulp meal doesn't cost more
Skills and roles change to suit and meet the demands of the world as required.
I still think that there's an awful lot of hot air about what AI is bringing and we're a number of years down the line in terms of it's appearance in mainstream businesses, in any real meaningful way.
Over the last 5 years I've seen more focus on making sure AI doesn't drop you in the shite, than I have about replacing the workforce of secretarial pools that are still manually typing away on their analogue type writers.....
So much scaremongering on a technology that has yet to prove its worth in mainstream business, other than making loads of money for power companies who are only too happy to sling mega/gigawatts of power to your bit barn.
This post has been deleted by its author
AI is exactly as good at physical jobs as it is at clerical and organisational ones, you just don't appreciate how bad it was at the latter until you need the output of its work.
This does entirely validate the belief that AI doesn't need to be good at anything to have a vast impact on the jobs market, it just needs salespeople who can convince the bosses it will justify firing people.
And of course, the gullible rubes who fall for the line will get their bonuses and move on long before the impact of their decisions finishes undermining the company, so there will be no meaningful consequences for anybody involved except the people who were good at their jobs and got fired.
This post has been deleted by its author
I see a lot of applications of AI for work companies do not want to do anyway, like customer support after the sale has been completed. Or at least, I see advertisement for such applications.
Currently, actual phone numbers for customer support are well hidden and you wait ages for someone to actually pick up the phone. In future, you might be finding these phone numbers easily and they will pick up the phone immediately. But there will be an AI at the other end of the line that reproduces the FAQ* in speech. All bets are off if your question in not in the FAQ. If the answers are bad, that is the customers problem, just as the non-answers of the human customer support persons are not their employers problem.
That is, AI will give the impression there is someone listening to you and solving your problems. Keeping up the impression is all that counts.
* AI is good at statistics. Statistics are good for Frequently Asked Questions. Much less so for Rarely Asked Questions (RAQ). But RAQ is why you called customer support in the first place.
The human version of chatbots - outsourced, barely trained call center staff - are just as bad. Just this week I had one company's representative swear up and down that they do not, under any circumstances, send postal mail to nonmembers, and therefore the postcard I got (with their company name, correct phone number, correct address, correct website...) must be a scam.
This post has been deleted by its author
This post has been deleted by its author
... until you need accurate output.
Using Google Translate, inputting the Japanese hiragana for "drug addict" returned the English phrase, "happy medicine cow".
Generative AI and large language models are not as useful/effective as their promoters represent them as being.
This post has been deleted by its author
This post has been deleted by its author
I really want to know how this is going to work. So, you've read so many of these reports by Gartnerites and Davosites that you've been brainwashed into believing that you really can just replace huge swathes of staff with this magic thing called AI. Off you go to the jobs fair and hire Sam Jungman who has just graduated with a shiny new degree in "AI". You sit him down at his new desk and say "right young Sam, your first job is to replace our accounts department with the new magic AI, and get it to automate our entire company's financial processes".
At which point I can imagine that all young Sam can say is "Oh shit, how do I do this?".
Yes, we've all played around with LLMs. If Sam asks "how to automate our accounts department", the answer will be wrong. LLMs are not good at that, and the question is way too generic anyway. If Sam is dumb, he'll try to implement the answer, and waste a whole lot of time and resources.
If he's smart, he'll gather knowledge from the field, learn company processes, ask a large number of much more accurate questions, learn enough to become able to tell when the answers are wrong, spend a lot of time doing just that, make action proposals, implement them, make sure they work, and, generally speaking, do roughly the same things a traditional software engineer does, only in a more convoluted way. At no point of that process has the "AI" done anything of value.
If Sam is smart he'll get the AI to generate a Powerpoint about "Using AI to leverage dynamic synergies" and when he presents this he will be promoted to "AI Grand Leverager of Synergies" and won't have to worry about the accounts dept
This post has been deleted by its author
There is a secondary component apart from the sales droids and marketing gurus efforts.
I recently had a conversation with an opposite number at a different company who was espousing how fantastic the new AI powered report generation was. What used to take a team of associates several hours per report was now generated in minutes!
Smelling a malodourous waft I spoke to the compliance person involved and they reported that they now had moved from 4 eyes review of the report to 8 eyes due to the errors being found, the people who knew the data and topic were no longer as involved in the production of the reports and people were assuming "I push the magic button and the report is spat out and hey presto it must be right the system said so".
If you think hallucinating AI is bad with somewhat curated data wait until you see what it does when fed a diet of not so curated and clean data.
So after investing considerable budget, personal political capital my opposite number was now trotting out the line about how fantastic his vision and execution was, the company is now AI enabled, money and time saved, etc. etc. This then gets repeated at board level and exec level to other boards and execs who, to be fair rightly, say hey! how come they have this awesome AI doodad that's saving the hundreds of hours a month and we don't?
So between sales/marketing on the front end and senior management not admitting any faults on the back end you end up getting spit roasted by the AI hype and spend inordinate amount of time educating people about what's real and what's hype.
But it gets worse, as anyone who has dealt with senior management and exec will attest to, there is a lot of wishful thinking going on too.
I wish there was some technology thing that increased our output for less cost
I wish there was some technology thing that figured this stuff out for me so I don't have to
I wish there was some technology thing that made our business process work properly without me having to actually sit down and think about the right way to do it
I wish there was a magic wand that just did the grunt work for me.
Ooooh this AI stuff sounds perfect, Hey!! IT bods, why can't we use this AI stuff...........
took a stab at it.......
The AI initiative
.
In the beginning, there was an AI initiative,
And then came the assumptions,
And the assumptions were without form,
And the AI initiative without substance,
.
And the darkness was upon the face of the workers,
And they spoke among themselves saying,
"It is a crock of shit and it stinks."
.
And the workers went unto their Supervisors and said,
"It is a pile of dung, and we cannot live with the smell."
.
And the Supervisors went unto their Managers saying,
"It is a container of excrement, and it is very strong,
Such that none may abide by it."
.
And the Managers went unto their Directors saying,
"It is a vessel of fertilizer, and none may abide by its strength."
.
And the Directors spoke among themselves saying to one another,
"It contains that which aids plants growth, and it is very strong."
.
And the Directors went to the Vice Presidents saying unto them,
"It promotes growth, and it is very powerful."
.
And the Vice Presidents went to the President, saying unto him,
"This new AI initiative will actively promote the growth and vigor
Of the company with very powerful effects."
.
And the President looked upon the AI initiative
And saw that it was good,
And the AI initiative became Policy.
.
And this, my friend, is how shit happens.
This post has been deleted by its author
This post has been deleted by its author
This post has been deleted by its author
I don't think we should under estimate the impact of AI on physical labour IF it really does start replacing 'soft' jobs. Once its smart enough to do paperwork jobs its just a matter of hardware.
As soon as there's robots that are sophisticated enough to perform $task then economies* of scale will do the rest. Hell, we can already automate all sorts of tasks using existing tech in my estimation.
Drones == delivery drivers replaced (hell, I might even get my food delivered to my door rather than almost delivered, only making it as far as the street outside).
Any number of existing home appliances can be automated, roomba anyone? You think it'd be hard to figure out loading and unloading a washing machine with an automated laundry basket? Emptying the trash? Run a carpet cleaner/floor polisher? Easy peasy. Bye bye cleaners and janitors.
Any kind of driving related job. 100% replaceable with automated driving, if they ever sort out the kinks, or just ban human drivers.
50 years or perhaps a little less and I reckon it could get there. There would of course be massive social upheaval.
*Assuming there's enough income left in society to pay for all these wonderful innovations har har.
Agree almost completely. AI development companies are already doing deals with robotics development companies, to combine the two.
As skills are honed by real world experience they can be pushed to any AI-enabled device, which is where the business model probably anticipates revenue streams from subscription services (remember BMW and seat warming subscriptions? It was shamed out of existence but it's inevitable).
3D printing of just about everything - including buildings (which look pretty unfinished these days, but that will improve), food, biological tissues, you name it, someone is working on it - will impact broad swaths of currently human-reliant jobs.
IMHO the next 2-3 years will see upheavals the like of which we have never seen before, and hardly anyone is prepared, least of all legislators.
Universal Basic Income is a must, as is Universal Healthcare, but authoritarian regimes will never implement those. The result is probably going to be massive street protests, and concomitant carnage in some countries.
Inevitably things will get worse before there is even a chance that things will improve - and by "things" I'm thinking of Life in general.
IMHO the threat to humanity is not so much AI and robotics, but that section of humanity that steadfastly refuses to believe there is any threat at all.
As for warfare, check out the Hellhive and drones. If Ukraine and Russia can do this, so can well-motivated terrorist groups, whether domestic or located on another continent.
with an automated laundry basket? "
I kind of do think that.
Because people have been talking about "The robot butler" since the 1960's
Seen one yet?
The thing about successful autonomous robot applications is their environment is either very structured (robot work cell) or very empty (drones flying in the sky)
But houses (especially British houses, with no minimum legal limits on floor plan) are anything but empty.
And robot manipulators are pretty bad at handling things that are floppy which is a pretty good definition of all clothes.
BTW it can be argued that a washing machine is a robot. It senses and manipulates its environment IE it's wash tub and produced a processed result.
Oh, is that all?
I mean, if you want Artificial General Intelligence it's just a matter of software.
Any twat can say "oh that's just....". It just means that you don't have a fucking clue what you are talking about.
"I mean, if you want Artificial General Intelligence it's just a matter of software."
Implicitly "how hard can that be?"
"If you want a multitasking operating system with a graphical user interface it's just a matter of software.".
After nearly four decades Windows we wouldn't be too confident in emphasizing the "just" when it comes to software.
Software is hard. Hard to get right - correct and complete without even considering resilience.
Unfortunately as a society we have been for decades degrading the roles that were assigned the tasks which are now slated for AI/LLM system. From phone menu systems, outsourcing the roles to godforsakenparts and scripted responses, the declining service standards (or lack of) have now been accepted as the norm so any AI replacements have been set a very low bar.
farmworkers, delivery drivers, construction workers, sales people, and food processing workers.
There are robotic arms, robotic eyes (i.e. cameras), voice recognition, voice synthesis, and more. It's only the robotic brain that's missing. If A.I. was actually as adaptable as a human intellect, it could interact with the world and quickly figure out how to do such menial labor.
Of course it's not... L.L.M.'s are basically at the level of incompetent sociopathic interns, who will answer your questions with the first results they find searching on Google, and insist it is correct, no matter how many times you tell them its wrong, and never learn, anything.
Was it perversity that the first vocation that popped into my mind on seeing the headline was that of a fluffer the existence of, Wikipedia assures me, is rapidly becoming apocryphal* due to the advances in technology.
I unconsciously discounted tosser as an occupation which for the "hands on" aspect would have to be first cab off the rank although the C-suite and WEF† participants strongly suggest I was premature in my unconsciously discarding this calling.
Actually the "adult entertainment" is one of the few industries that easily benefits from contemporary AI/LLM - characters with distorted anatomy, magically acquiring and discarding extra appendages and limbs, should keep the audience experience fresh for years to come.
* in that those in the industry (fluffees?) disclaim their ever existing.
† sinister tossers.
RE. construction work and laborers
Oil sands in Northern Alberta pays drivers a lot of money to drive 400 ton dump trucks
https://www.cat.com/en_US/products/new/equipment/off-highway-trucks/mining-trucks/18093014.html
Oil companies are doing lots of work using automation and AI to replace the drivers where a well mapped route exists and the machine goes from the bucket machine to the processing plant continuously day in day out. And its pretty good in that controlled environment, never speeds, always reverses into the same exact distance from the machine even when the machine moves a few feet to access new material. uses the engine breaking where possible ( or what ever the diesel electric equivalent is) and generally runs the machines in a more controlled and mechanically sympathetic fashion than human drivers.
Now the guy running the bucket is still needed for now because reading the ground or maximizing the fullness of each scoop is a bit of an art, but I wouldn't be holding my breath for a pension out of it.
https://www.cat.com/en_US/products/new/equipment/hydraulic-mining-shovels/hydraulic-mining-shovels/102960.html
In situations like those where the time and motion bods get involved, you can be very sure someone with an abacus is counting the pennies and coming to a bad (for the workers) conclusion.
But none of this is new, French workers went on strike and ransacked and set fire to Jacquard looms arguably the first programmable machines to exist (sorry Babbage, they were there before you) because they were being displaced and wages sank. Modern ships have tiny crews running gigantic behemoths.
The difference this time, I think, is those were all examples of automation and could be proven through logic and empirical testing to deliver a "correct" result. For now AI is very much a black box, despite chain of thought structures being put in place. But that too will pass. And then it's back to the age old cost versus profit. And workers are not cheap.
Anyone in the trades is probably safe for a few decades, electricians, plumbers etc. Nurses will find a fair bit of stickiness in their jobs even if humanoid robots with nursing skills become adopted, who wants that when you are in pain/dying. If anything I think white collar jobs are far more at risk particularly if it is a rote repetitive job.
On the flip side any job requiring flair, design, artistry, critical thinking, creative, or imagination will always be needed. Unfortunately there is a limited number of those that pay well, starving poet never appealed to me as a career.
Just so it's not all doom and gloom, I will posit an upside.
There is a fair argument that some of the most amazing advances in thought, mathematics and sciences occurred during periods of times when gifted thinkers had time to think and mull. They had slaves to do the menial and manual work. I'm thinking Egypt, Mesopotamian Babylon etc. Or maybe we all are slaves now and the future AIs will dream of electric eurekas.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
