AI skills shortage more than doubles for UK tech leaders The number of UK tech leaders reporting a dearth in AI skills has more than doubled in the last year, according to research. A survey by tech recruitment specialist Harvey Nash found that around half UK tech leaders (52 percent) say they are suffering an AI skills gap in the past year, compared with 20 percent who said the …
This post has been deleted by its author
Exactly. Employers still don’t realise how deep the hole is. A young person today can either take on crippling uni debt, spend their free time staying ahead of a tech treadmill, and shoulder all the pressure - or do a few months’ training and earn the same (or more) laying bricks, wiring houses, or driving vans.
Agreed. I wish employers woulds stop referring to "not cheap enough" as a "skills shortage"...I also wish they would stop offering equity in lieu of payment as well...every man plus his dog knows this scam now and has done for decades.
The UK has large number of technical, talented individuals...we're known worldwide for it...our technical skill is respected the world over...it is widely acknowledged that without Britain and it's engineers throughout history there is no "modern world"...and yet, at home our own fucking people, our own employers think we're worth jack shit because they can hire Eastern Europeans, Indians and Asians for peanuts.
Listen up employers, because here is how it is...if you want to build a product that is respected and could someday be looked back upon as a turning point for a given industry, the beginning of a new market, a product with global significance...you hire British engineers and you pay them well...
If you want to produce the next wanky ad supported phone app that nobody will have heard of and definitely won't remember in 18 months, you go for the cheap talent. If you want to build something market leading, you hire British talent...It doesn't matter how amazing "your idea" is, you are nothing without a solid team of engineers.
Personally, I am fucking sick of being brought in at the last minute to bail out failing projects after the budget has been pissed up the wall and being the lowest paid guy in the room.
The maths is insanely simple. If you pay the rate for British engineers, you get far more than you pay for. British Engineers are far more than just folks that bang out your idea. You're hiring a partner essentially, they will go to lengths to understand you, your product, your target market and your goals, they will understand your vision and deliver it. If you go with cheap backwater engineers, you get less than you pay for. They don't give a shit about you, your vision, your market or your goals...they only give a shit about delivering just enough to get you to sign off so they can raise an invoice.
For clarity for the numbskull, loathesome, wanker executives and investors out there in Britain, this is how you approach rewarding your engineers and hiring them:
1. If you are a taking a salary, I will not accept equity in lieu of payment.
2. If you haven't allocated all of the seed capital towards building / growing the business, I will not accept equity in lieu of payment.
3. If you have traded large equitable stakes in your company to investors in return for fuck all money, I will not even listen to you.
4. I don't care how good your idea is, it's worth nothing until it is a working product.
5. Whatever I build is mine until you pay for it. Either pay well, or pay up front.
6. Yes, yes...your idea is a million pound idea...but have you spent any time building a spec? A plan? Research? No? Fuck off then.
7. Do you have a clear breakdown of how seed capital is going to be spent? No? You're a fiddling bastard...get out.
8. Do you have more executives, paid expert consultants and board members than techies? Yes? Did you give them equity as well for nothing? This is why you have no money to spend, fire them...you don't need them.
9. Do you have £500k+ in seed capital but somehow you want to build your prototype for less than £10k? Just fuck off.
10. Did you hire graphic designers before you hired engineers? Yes? Fuck off.
Nice speech but underneath you already know the answers.
"if you want to build a product that is respected and could someday be looked back upon as a turning point for a given industry, the beginning of a new market, a product with global significance...you hire British engineers and you pay them well...If you want to produce the next wanky ad supported phone app that nobody will have heard of and definitely won't remember in 18 months, you go for the cheap talent. "
Sorry to say that your wanky execs aren't asking that questions, they want to know what will produce them the most profit and therefore the biggest bonus in the next 2-4 quarters. Quality, functionality, longevity are all of zero importance. AI is a just the current buzzword term to get shit-talkers up the greasy pole a little higher.
Man, stop being ungrateful...everyone knows "the idea" is the most valuable part of any business. It takes a lot more time and skill to come up with an idea...the development, execution, deployment, research, marketing, scaling and distribution as we know is worth fuck all.
Drawing a triangle on a sheet of papyrus was the real feat, not the generations of effort and hundreds of thousands of engineers, masons and craftsmen that built the pyramids...they were all scum. What side of history are you on?
Without the idea...nay...the vision, the sheer audacity to dream without boundaries the people involved are worth nothing.
/s
Why would they even want people skilled in AI, when the whole point is to fire people? They just need to fire them and the rest will have to take on the extra work and do it however they can, as always. If remaining employees have access to ChatGPT then the board can say their company is an AI company.
Because you need the AI guy to observe the workforce in order to train the AI to replace the workforce. Which is an extremely difficult job, because most people think they're unique snowflakes and they can't be replicated...and yet, their job is 90% reading and writing emails, 5% googling and 5% presenting powerpoints. I hear garbage all the time from people related to this..."it's difficult to explain my job, it's very nuanced", "but what about the human element?" etc etc...the idea behind AI being integrated into a business isn't necessarily to replace humans, you need them for "the human element". That's the one thing a human brings to the table that AI can't...being human...but what AI can do, if it is properly developed, is take a ton of the load away on repetitive tasks...like processing data, producing reports, interpreting data, looking for patterns etc etc.
The biggest threat to people with regards to AI is not that they might lose their jobs, it's that their work overhead will be reduced and the actually useful part of their skillset will shrink...which means you spend less time working...in a perfect world, you'd be paid the same even though you're doing less and you'd have more free time, the unique bits of your skillset become worth more, because you can spend more time focusing on that rather than being drowned in repetitive bullshit that anyone or anything can do...the reality is, you'll simply be paid less because in the eyes of the business, you're doing less and therefore providing less value.
AI, despite the recent explosion in LLMs etc, is not an entirely new field...quite a lot of people in tech have a background in AI...they've just never used the skills because it's never been in demand before.
I've been working with AI for almost 8 years now...building systems that can learn and interpret data and perform actions...it's not actually that difficult because a lot of jobs are repetitive, due to human nature, we like to make things repeatable to make them easier...the hardest part is sourcing / collecting the relevant data, especially from fellow humans, because they won't recognise that large parts of their job are repeatable and don't need to be done by a human...outside of the obvious data (i.e. personal data), mankind hasn't really been recording things in great detail for very long...so it's pretty hard to find / build useful datasets in order to make something AI based useful outside of language models and/or image diffusers...because we have tons of images and texts to use for that...but we don't have, say, a few centuries of weather measurements...or road traffic data...MRI scans...xrays...etc etc...the historical data there is relatively small...this is the sort of stuff you need data scientists for...which absolutely none, literally zero, nada...bugger all...organisations I've worked with to build AI have had.
Because one person skilled in AI can observe an entire department then replace them, and train more folks to do the same, up to a maximum number of 6-8 (depending on whether you need reserves for your eSports clan). Starting at the bottom and working all the way up to the CEO...then eventually it will just be all AI and techies. Which to me, as a techie, sounds perfect.
I'd love to provide support for businesses with no humans other than techies in it with AI looking after me. It'd be great.
What a future...just AI run businesses, eSports clans for the guys that look after the AI and billions of people doom scrolling TikTok buying shit to fund the AI.
Ah, but how do the serfs earn money? I hear you cry! That's simple, their entire existence will be just creating content for TikTok for which they get paid based on views. TikTok becomes a game of "go viral or die".
The premier events for the year will be AI engineer eSports clans facing off against poor skinny TikTok influencers, Roman Gladiator style. A fight to the deletion (of your social profile and monetisation status). Participants will be chosen based on their "edginess" and for producing content that is just shit.
https://imgur.com/a/CmUMFOh
Dont worry,classrooms full of people that have memorized all the buzzwords are being produced daily. They will flood the industry and sit there doing nothing most of the day,slowly learning from mistakes.
Uk guys will continue to walk dogs and be oblivious to all the jobs they are missing out on.
Time to go back to aptitude tests for jobs for UK school leavers with good exam passes and character. Even someone with no techno bollox qualifications is probably better than a lot of people getting hired.
They can be put through a 12 week training program and due to their 'aptitude' be a functional employee.
Same for most jobs. Bin off the certification culture and bits of paper people wave in your face.
It's because certs like the MCSE are widely misunderstood. The suggested requirement for sitting the exams used to be 2 years of prior experience...basically the certs were supposed to represent at least 2 years experience...but the whole thing got subverted when recruiters started latching on and we started getting bootcamps and other shit like that...training services that "guaranteed a job" as long as you paid £20k for the course etc...
It might surprise you to know, that current Comp Sci degrees follow a curriculum that is almost identical to an MCSE with a few extra electives.
My nephew recently completed his degree and I had a look at the course materials...it was like MCSE-Lite...it was shocking. I go all the way back to Windows 2000 certs, and I remember "The Beast" as some people called it...the 70-216...a modern day comp sci degree seems to follow a similar path to the old "2000" MCSE but without "The Beast".
I don't think anyone in the industry considers MCSE level certs to be any indicator of skill, certainly these days...me and a friend of mine used to compete on MCP exams for fun, we were both in our late teens and unemployed and the dole funded them at the time. So we used to pick one at random, book it for a week later in the late afternoon, turn up to wherever the test centre was a few hours early. Play some pool and get trollied...then sit the exams...they were that easy to pass...we'd both been colossal nerds since our early teens, so it wasn't like we went in with zero knowledge, but we certainly didn't have to prep for them. My mate had one of his results rejected because he managed to get a perfect score, which apparently isn't possible on those exams unless you cheat. He came out of the exam into the waiting area and said to me "I think there was a fake question in my exam, there were 3 right answers and a wrong one, in the 3 right answers there was one that was really basic so I picked it"...dude came out with the scores a short while later, I scored like 92%, but he scored 100% and his print out said "subject to review"...he got a call on the way home saying his result was rejected and that he had to wait 2-3 months to sit another exam...which is where we stopped...we had 14-15 MCPs each at that point. We'd been doing 1-2 a month for about 9 months...we moved onto Cisco stuff after that, both got the CCNA and a bunch of other individual modules.
I did eventually do more MCPs further down the line when my employer, for some reason despite me having more MCPs than fingers and toes, requested that I do more...so I did some of the fringe ones that I hadn't done like Sharepoint, WSUS etc etc...those were tricky but purely because they were extremely boring.
Usually exactly the type of skill needed to make "AI", that was sold to that specific CxO as a magic cure for all the problems of the company, work. Easy, right?
Try to differentiate between machine learning, generative AI, or maybe neuronal nets - and you are out for not being enough of a magician being able to work with unicorns and pixie dust :)
The current AI climate reminds me a bit about the time leading up to the dotcom crash: Companies were crazy to hire whoever could pass an interview claiming to be a "programmer" to build the next big dot.com thing, without being able to verify the claims of the "programmer" (and mostly not even interested in clarifying ...).
However, the dotcom boom was still - in a technical sense - more sane than the current AI boom.
On a more serious note:
Managing expectations, keeping customers and managers grounded, understanding the basics and knowing the limitations should get you going for any serious AI effort - and it will keep you from joining the pure hype ones.
Another year, another breathless headline about a "skills shortage" - otherwise known as companies refusing to pay market rates for highly specialised work. There’s no actual lack of AI talent - just a shortage of mugs willing to spend years mastering complex systems only to be underpaid so some executive can afford a third holiday home or a new yacht.
The so-called "gap" only appears when employers filter out anyone who expects a decent salary, job security, or the ability to work on something that isn’t doomed to become next year’s "strategic pivot."
If they really wanted to attract skilled people, they’d stop demanding five years’ experience in tools that came out last Tuesday, and maybe stop treating AI engineers like disposable buzzword dispensers.
"pay market rates for highly specialised work"
I sort of agree with that. Tech used to be a generalist profession. You may have had a specialism, but you also had a broad set of skills outside that specialism. The problem with "skill shortages" isn't really a lack of people with skills, it's a lack of people with more than one skill.
Look at the games industry for example, they have compartmentalised to fuck and it's causing endless shitty games to be made. You have hundreds or thousands of people working on a AAA game, and it ends up being a janky, badly optimised and incoherent mess...on the flipside you have small teams of indie devs (usually around a handful of people) who are generalists and they are putting out some amazing games.
Look at firms like ID, Raven etc etc...they were tiny little teams of generalists in the 90s...they put out some amazing games back then...they have since turned into massive sprawling behemoths and they're putting out shit games.
Specialists. Kill. Tech...specialists need to fuck off back to McDonalds and stand on their chicken nugget station...you can't be a specialist engineer in terms of the skills you possess. For example, it's not possible to be a bolt tightening specialist, that's stupid. You can only be a specialist in the things you build. Like a bridge building specialist. Within that level of specialism, your skill set is generalised. You know how to attach this to that, how this material works under that load and how this can work with that etc etc...there is no "Hey, it's not my fault, all I do is tighten the bolts, I don't drill the holes mate"...that's not engineering.
When you have a house made up entirely of specialists, responsibility is so diluted, communication is so fragmented and pipelines are so convoluted that it isn't possible to build anything with any level of coherence, all you can achieve with this sort of setup is the speed you shit out slop...and no single person can be blamed...you end up with an army of jobsworths.
Tech is not fast food, you can't teach individuals singlar aspects of a process (you tighten the bolts, you drill the holes, you carry the planks between the guy that drills the holes and the guy that tightens the bolts) and expect to get good results.
Specialists get generally shitty pay for a reason, they're only good for one thing. Generalists are the ones that make the big bucks. Sure a generalist might not be able to drill holes as well as a guy that has only drilled holes for 20 years, but the hole drilling guy hasn't a fucking clue how to apply his skill in a wider context than just drilling where he's told to...he can't build a product, he can't solve a problem...he can only drill holes...it's all he's ever known, will know and all he can do and if the company decides to stop using bolts and moves to rivets instead, he's fucked.
If you decide to be a specialist in tech, medium to long term, you are fucked.
I am a generalist in tech, always have been...I have never been out of work since I was 19 years old...which was 22 years ago. Only when I started out as a junior was my pay crap...as is expected when you're a junior in your first couple of years. I know many specialists, folks I went to college with etc etc...and they have been in and out of work almost constantly and they have never been paid well.
I command higher rates because my customers get more out of me that they would out of a specialist. Value for money. Instead of assisting with one aspect of a system, I can assist with the whole damned stack. Top to bottom.
I know specialists get butthurt because they pour a lot of time and effort into being as good as they can be in their narrow field, but these are just the facts.
You only ever need specialists for short periods of time, to come in and look at the work that has been done and suggest improvements in their specific area of expertise. That's it...you don't get specialists to build shit. Ever. Because you end up needing hundreds of them and they don't give a shit about the big picture, they only care about their narrow field of expertise and doing the best they can there...they can't even see the bigger picture.
They are desperate to hire people who can take the magic snake oil currently known as 'AI' and replace x% of the workforce with it.
They should go the full Monty and set up a unit to turn lead into gold whilst they are at it.
And if they manage to create purest green, Lord Percy style, they can bank serious cash from a government eco credits scam too.
Sensible people are doubling down on their internet security, using less, simpler tech, paying for competent staff, reducing non-essential expenditure and focusing upon their core markets to survive the global financial disaster on the horizon.
Avoid AI. If you use Windows, switch to LTSC.
It's time to get your free ad supported web based native multiplatform iphone android windows linux macos ESP32 ARM x86_64 blockchain NFT gamified web 2.0 data driven self help food diary banking social media gambling porno full stack progressive sharding responsive microservice REST scaleable GPU powered low power solar powered wind farm nuclear high performance crypto cybersecurity big data exclusively for everyone freemium subscription loot crate microtransaction enshittified DNA ancestry short form video tweet based WEB3 referral link merch supported youtuber backed twitch stream multifactor VPN offshore cloud based todo list app AI'd up the wazoo with large language copilot diffusion deepseek GPT llamas my friend. Haven't you heard?
Full stack is a journey not a destination.
This post has been deleted by its author
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
