Bottom Line Up Front

Every company in the West is about to become a different company. Not in a decade. In months. The only question left is who moves first: the board, the market, or the mob. This week Microsoft picked one answer. The prediction markets picked another. Sam Altman’s front gate picked a third. A girl expelled from school at fifteen, now standing in a barrister’s robes, picked a fourth. Most leaders will misread all of them.

Subscribe now

1. The Zombies on Your Payroll

Peter Drucker asked one question most managers spend their careers avoiding. If we were not already doing this, knowing what we now know, would we start today?

Try it on one of your own departments. The honest answer is almost always no.

Drucker called the consequence “systematic abandonment”. In plain English, it means stopping the activities that once made sense, still draw salary, and quietly crowd out whatever comes next.

For most of the last century, leaders could dodge that question for years. What they knew changed slowly. Strategy could wait.

That world has gone.

What we know now changes between one frontier model release and the next. Decisions Drucker imagined being taken every few years are now being forced on companies in months.

A clarification, because it matters. By zombies, I do not mean people. I mean functions, layers and workflows built for yesterday and never redesigned. The middle manager whose role is to summarise reports for the level above. The centre of excellence that exists mainly to explain itself. The team producing drafts that a model can now generate in seconds.

Zombies are not your colleagues. They are the structure around them.

And that structure is starting to break.

The evidence is arriving first from America’s corporate giants. Oracle has begun cutting 30,000 roles while investing heavily in AI infrastructure. UPS plans to shed up to 30,000 operational jobs as it consolidates into fewer, more automated hubs.

The causes differ. AI in one case. Automation in another. A changing customer mix in a third. What unites them is Drucker’s question.

Not one of those companies would build itself today exactly as it is.

Here is the uncomfortable truth. If I were running a 200-person company again, and I knew the real shape of the business was 25 to 50 people smaller, could I make those cuts? I do not know.

I have hired people across the companies I have built. I know what a job means to a mortgage, a family, a life. Leaders tell themselves people can be redeployed or retrained. Sometimes they can. Often, at the speed the market now demands, they cannot.

That is the part Drucker never quite captured. Abandonment is not a spreadsheet exercise. It is grief, entered into knowingly, because the cost of refusing is greater.

Most leaders cannot do it until the market does it for them. That is how you become Kodak, protecting film long after the digital future was visible. That is how you become Nokia, defending an ageing platform while a new one takes the world.

So here is the question for every chief executive. If you were founding your company this week, with a credit card and an API key, would you build what you have now?

If the answer is no, the decision is already made. The only question is whether you act on it, or wait for the market to act for you.

2. Microsoft Breaks Fifty-One Years of Silence

This week Microsoft did something it has never done in its history.

On Thursday the world’s most valuable software company offered voluntary retirement to roughly 8,750 American employees, about 7 per cent of its US workforce. It is the first such programme in the company’s 51-year life.

The eligibility rule is called the Rule of 70. An employee’s age plus their years of service must equal 70 or more. A 50-year-old with 20 years at the company qualifies. So does a 45-year-old with 25 years of service. Details land on 7 May. Eligible staff have 30 days to decide.

This is not a layoff. It is a buyout dressed as a benefit. And it is the clearest signal yet of what the AI transition actually looks like inside the world’s largest technology companies.

Microsoft reported 38.5 billion US dollars in net income last quarter. It is projected to spend between 110 and 120 billion US dollars on AI infrastructure this fiscal year. The company is not cutting because it cannot afford its people. It is cutting because it has chosen to afford something else.

Chief People Officer Amy Coleman’s internal memo framed the programme as giving eligible workers “the choice to take that next step on their own terms”. Read that carefully. The choice being offered is not whether to leave. It is whether to leave now, with dignity and a cheque, or later, with neither.

The Rule of 70 is not random. It targets long-tenured employees whose institutional knowledge was built around the pre-AI Microsoft. The Windows XP generation. The Server and Tools expansion. The early Azure buildout. These are the people whose expertise centres on systems Microsoft is now automating or deprecating.

Microsoft is not alone. Meta is cutting 8,000 jobs on 20 May. Oracle cut up to 30,000 in March. Amazon has signalled roughly 30,000 cuts across AWS and Alexa in the first half of the year. Meta’s capital expenditure on AI infrastructure is projected to reach 135 billion US dollars.

The arithmetic on Wall Street is now ironclad. You cannot fund a hundred-billion-dollar compute pivot while carrying pandemic-era headcount. Software engineers are no longer the most valuable asset on a technology balance sheet. Compute is.

Microsoft’s version is kinder than Meta’s, because Microsoft can afford to be. The financial payout, the extended healthcare, the 30-day window. All of it is designed to minimise legal risk and reputational damage. It is a polite exit.

The message is not polite. The veteran workforce that built the pre-AI Microsoft is being offered a seat in the lifeboat while the company sails into a different century.

This is not a transition. It is a replacement, conducted with paperwork and a 30-day window. Drucker would have recognised it instantly.

3. What the Betting Market Knows

The sharpest read on what is coming next is not in any analyst note or McKinsey deck. It is on Polymarket.

The contract on whether US technology layoffs in 2026 will exceed the 2025 baseline of 447,000 currently trades at a 90 per cent probability. Real money. Real traders. Real consensus. The sister contract on Q1 alone has already moved to 98 per cent. The market opened in March. It has barely wavered since.

Nobody is paid to be wrong on Polymarket. Strangers with no equity in any of these companies and no axe to grind are wagering their own wallets that this is the year the dam breaks.

The market has already made the decision that boards are still pretending to debate.

When the boardroom, the betting market and the street converge on the same outcome, only the timing is in dispute.

4. The Luddites Come for Altman

At four in the morning on 10 April, a 20-year-old man from Texas threw a Molotov cocktail at the front gate of Sam Altman’s house in San Francisco.

The gate caught fire. Nobody was injured. The attacker, Daniel Moreno-Gama, was arrested an hour later at OpenAI’s headquarters, carrying a jug of kerosene, a lighter, and a document prosecutors say outlined his opposition to AI and what he called humanity’s “impending extinction”. He had published similar views on a personal Substack. He is charged with attempted murder.

Three days later gunfire was reported near Altman’s home and two more arrests followed. Earlier this month someone fired thirteen rounds at the home of an Indianapolis city councillor who had voiced support for a local data centre. The note left behind read: “No data centres.”

This is the backdrop to everything else in this issue. It is also not new.

In 1811 English textile workers began breaking into factories and smashing the wide-frame looms they believed were destroying their livelihoods. They called themselves Luddites, after the mythical General Ned Ludd. Within five years the British government had made frame-breaking a capital offence.

The Luddites were not anti-technology. They were anti-accountability. They were fine with the looms. They were not fine with a generation of factory owners using those looms to bypass labour protections and drive wages to subsistence while getting rich. Parliament ignored their petitions. So they picked up hammers.

The pattern repeated a century later in France, where disgruntled workers supposedly threw their wooden shoes, or sabots, into the gears of new machinery. That is where the word sabotage comes from.

The template is consistent across two centuries. When people feel a technology is being used on them rather than for them, the response moves from vocal to physical. The violence is almost never about the technology itself. It is about the power dynamic the technology creates.

History teaches four lessons, and every boardroom should read them now.

Disruption without a safety net radicalises people. The Luddites did not reach for hammers the day the loom arrived. They reached for them after Parliament ignored seven years of petitions for transition assistance. Indifference is combustible.

Consent is the only durable business model. Apple scans your face, but gives you a lifetime of searchable photographs. Google tracks your location, but ensures you are never lost. Ambient, unavoidable technology breeds resentment. AI is currently being installed into daily life with neither opt-in nor recourse.

Transparency prevents conspiracy. When a technology is a black box, the public fills the void with its worst fears. Even the builders of the large language models cannot fully explain their internals. Into that gap walks every manifesto-writing 20-year-old with a grievance and a bottle of petrol.

Regulation is a pressure valve, not a burden. When consumers can sue, when governments can fine, when products can be recalled, anger has somewhere to go that is not a billionaire’s front gate. Remove those options and the only recourse left is the street.

Aaron Zamost argued in the New York Times this week that Americans now feel about Big Tech what they feel about health insurance companies. They cannot opt out, cannot switch, and cannot hold anyone to account. Seventy-seven per cent of Americans believe AI could pose a threat to humanity, an idea Altman himself has publicly advanced.

Altman will survive this. OpenAI will survive this. The industry should not act surprised.

The attack on his front gate was not a watershed. It was a warning shot.

5. What the System Missed

This week I wrote about Leonie Hughes.

Expelled from school at fifteen. A mother, in effect, to her younger sister at eleven. She grew up surrounded by domestic violence and substance abuse. The education system looked at her and decided she did not fit. Today she is a barrister. She chose law, she says, because it was one of the hardest careers to go into, and she wanted to prove the doubters wrong.

Her story is not an inspirational anecdote. It is an indictment of how we measure potential.

We have built an education system that rewards compliance over resilience. If you thrive within its rules, you are bright. If you struggle, you are sidelined. The children who fall out of the back of that system are not the ones lacking ability. They are often the ones carrying more than the system knows how to see.

In an economy about to be reshaped by machines that can pass exams better than most humans, this matters enormously.

AI can summarise. AI can draft. AI can predict. AI cannot know what it is to raise your sister while still a child yourself. AI cannot decide to become a barrister specifically to embarrass the people who wrote you off. AI cannot draw on a life to know which lie in a witness box is the dangerous one.

The qualities AI cannot replicate are the qualities we spent a century teaching children to hide. Stubbornness. Irregularity. Lived friction. The kind of grit that gets you kicked out of a classroom at fifteen and into the Bar at thirty.

Most of the Leonies this country has produced are not barristers. They are on zero-hour contracts, dismissed at sixteen, told by a system built for a different century that they were never going to amount to much. A decade from now, when half of what we currently call white-collar work has been absorbed into a GPU cluster, we will realise we discarded exactly the people the new economy needed.

It is time to stop asking what is wrong with the child and start asking what is wrong with the system.

The Sunday Signal Tech & AI Layoff Tracker | Week 18 | 19–25 April 2026

The Signal: Meta’s 135 billion-dollar compute bloodletting

The “Year of Efficiency” was a warm-up. Welcome to the era of the Mega Layoff.

This week Mark Zuckerberg dropped the largest single technology bomb of Q2. Meta Platforms announced it is cutting 8,000 jobs, roughly 10 per cent of its global workforce, and eliminating another 6,000 open requisitions. The rolling cuts officially begin on 20 May.

If there was any lingering doubt about the AI Launder thesis, Meta has now codified it. This is not a macroeconomic adjustment. It is a brutal, direct capital reallocation. Meta is accelerating its AI infrastructure buildout, projecting 135 billion US dollars in capital expenditure to fund data centres and liquid-cooled compute clusters.

The maths on Wall Street is ironclad. You cannot fund a hundred-billion-dollar compute pivot while carrying pandemic-era headcount. Oracle traded 30,000 humans for Nvidia chips in Q1. Meta is now trading 8,000 humans to feed the Llama 4 infrastructure. Software engineers are no longer the most valuable asset on a tech balance sheet. Compute is.

The blue-collar pivot: the rise of the AI plumber

As Meta axes its white-collar software engineering layers, a paradox is emerging. Employment at US data centres has surged, and Meta itself just launched a free fibre technician training programme.

The new six-figure tech job does not require a computer science degree from Stanford. It requires a hard hat. Senior electricians specialising in liquid cooling and high-density fibre cabling are commanding salaries of 200,000 to 300,000 US dollars. The white-collar trade job is booming. The builders of the physical AI infrastructure are inheriting the earth while the middle managers of the SaaS era are handed severance packages.

Nike’s tech wipeout

The rot is extending aggressively beyond traditional Big Tech. On Thursday, 23 April, Nike confirmed the elimination of 1,400 jobs, roughly 2 per cent of its 76,000-person global workforce.

The critical detail is not the total number. It is the target. The vast majority of these cuts are centralised entirely within Nike’s technology department. Under new chief executive Elliott Hill’s “Win Now” turnaround strategy, legacy non-tech brands are abandoning the zero-interest-rate fever dream of operating like technology companies. Nike is gutting its digital overhead to return to the core mechanics of making and selling shoes.

The restructuring reality check: BBC and Lucid

Not every cut this week was a compute pivot. The traditional margin crush continues across legacy media and automotive.

• BBC (Media): 2,000 jobs cut, 10 per cent of the workforce. The biggest reduction in 15 years. A desperate attempt to find 500 million pounds in savings as traditional licence fee income craters.

• Lucid Group (EV/Auto): 319 jobs slashed at the Newark headquarters. The third straight year of cuts at that address. Part of a broader 12 per cent global reduction following a 2.7 billion US dollar net loss in 2025.

• PNC / FirstBank (Banking): 777 roles eliminated at FirstBank’s Colorado headquarters. A classic post-merger integration squeeze by PNC.

High-probability targets: Week 19

• Non-tech “tech” departments. Nike’s move is a massive leading indicator. Expect major retail, CPG and legacy apparel brands to begin aggressively pruning their bloated in-house technology, app and digital innovation teams. If off-the-shelf AI tools can run your e-commerce stack, you do not need a 500-person in-house engineering squad.

• Mid-tier social and ad-tech. With Meta setting the new efficiency baseline and weaponising its AI to monopolise advertising targeting, mid-tier social platforms and ad-tech vendors will be forced to slash headcount just to maintain acceptable operating margins.

🚀 Final Thought

Every boardroom will eventually answer Drucker’s question. Most will answer it too late, after the market has answered for them.

Microsoft has chosen a polite exit. Meta has chosen a brutal one. The prediction markets have priced in 90 per cent certainty. Sam Altman’s neighbours have installed extra security. A 20-year-old from Texas has chosen a Molotov cocktail. The same story, told in five registers.

The AI revolution will not be resisted in the ways its critics expect. It will be resisted by shareholders demanding the compute pivot. By traders pricing in the bloodletting. By employees refusing the package. By young men with manifestos. And by the Leonie Hugheses this country has already written off but who remain, somehow, the people the next decade needs most.

It is a power story. It always was.

This is not a transition. It is a replacement.

Most of them will realise this too late.

Until next Sunday, David

Bottom Line Up Front

History does not repeat itself. But it borrows the same costume every time. Last Friday I wrote in the Yorkshire Post about clearing out my garage and finding a Webvan stress ball, a Yahoo! mousepad, and an Ariba t-shirt. Those artefacts of the dot-com era cost investors hundreds of billions of pounds. The companies behind them died. The infrastructure beneath them did not. It became the plumbing for everything that followed.

The AI era is doing exactly the same thing. The pattern is identical. The displacement is real. The casualties are already bleeding. And an alarming number of fund managers and company executives are standing at the podium this earnings season with straight faces, telling investors that AI is helping their business. Most of them are wrong. Some of them know it.

This week, on the fiftieth issue of The Sunday Signal, we examine three things. First, what the sock puppet actually taught us, and why the lesson applies right now. Second, the full sweep of technology disruption across forty years, told honestly, without the survivor bias that distorts every boardroom presentation. Third, an investor’s guide with a concrete hot list: who is truly insulated, who is already dying, and why the most dangerous sentence in finance right now is “AI is helping everyone.”

1. The Sock Puppet Never Changes. Only the Sock Does.

In my column in the Yorkshire Post on Friday, I described unearthing a collection of dot-com relics from my garage. The Pets.com sock puppet. The Webvan stress ball. The Ariba t-shirt. I dropped the lot in the recycling bin. The lesson, I wrote, is not so easy to dispose of.

The Super Bowl in January 2000 remains the most instructive single moment in the history of technology investment. Fourteen separate dot-com companies each paid around two point two million dollars for thirty seconds of airtime. Pets.com sent a sock puppet. The company was bankrupt nine months later.

That game was not the peak of innovation. It was the peak of theatre. The revenues were not there. The margins were not there. In many cases every transaction lost money. The business model was not merely deferred; it had never existed.

Here is what people consistently misremember about that era. The collapse of Pets.com, Webvan, and Kozmo was not evidence that the internet was overhyped. It was evidence that speculative capital had attached itself to weak models. Amazon survived. Google survived. The fibre-optic cables that were laid at extraordinary cost during that period became the arterial infrastructure for the modern internet. The disruption was real. The casualties were selective.

There is a direct application to 2026.

Ask yourself this: of the companies you have invested in, partnered with, or work for, how many are the Pets.com of this cycle? They have raised money. They have a plausible story. They have a sock puppet, or the modern equivalent, which is a beautifully designed AI dashboard that wraps around a database and charges per seat per month.

The comparison I drew in the Yorkshire Post column is worth sharpening here. At its peak, Ariba was trading at over one hundred times its annual revenue, with investors paying for cash flows that never materialised. Cisco, the infrastructure darling of the era, peaked at twenty-six times sales. Now consider Anthropic: founded in 2021, already trending towards thirty billion dollars in annualised revenue, and trading at roughly thirteen times that figure. The multiple on genuine AI infrastructure is lower than the multiple on speculative dot-com theatre was at its peak. That is not a bubble. That is a structural repricing of computing power.

But Ariba is not the story. Ariba is the cautionary tale. The story is what survived. The entrepreneurs who built on that speculative infrastructure after the crash, buying assets at pennies in the pound, created the most valuable companies in human history. Around that same genuine foundation today, a second layer of companies has grown up that has no more durable a moat than Pets.com had. Their business model is not innovation. It is theatre.

Anthropic is not the risk. The risk is the company presenting to you this week with a ChatGPT wrapper and a go-to-market slide that describes itself as an AI-powered platform. Strip out the AI language and what you have left is a project management tool with a chatbot bolted on and a per-seat subscription model heading directly into the SaaSpocalypse.

The sock puppet is back. This time it is wearing a suit. And it is presenting at your next board meeting.

Subscribe now

2. The Pattern of Destruction Is Always the Same. Only the Technology Changes.

To understand who loses in the AI era, you have to understand who lost in every previous era, and why. The mechanism of displacement does not change. Only the underlying technology does.

The Mainframe-to-PC Era: The Unbundling of Power

In the early 1980s, Digital Equipment Corporation was the second-largest computer company in the world. It employed over one hundred thousand people. Its founder, Ken Olsen, had famously said in 1977 that there was no reason for any individual to have a computer in their home. By 1992 the company was haemorrhaging cash. By 1998 it had been absorbed by Compaq. Compaq itself was absorbed by Hewlett-Packard three years later.

Wang Laboratories was making a billion dollars a year selling word-processing systems to enterprise clients. It filed for bankruptcy in 1992. Prime Computer, Data General, and a dozen other mainframe-era giants either collapsed or shrank into irrelevance inside a decade.

The mechanism was simple. Microsoft and Intel decentralised computing. They standardised the components. They shifted power from hardware gatekeepers who charged enormous sums for vertically integrated systems to a software layer that sat on top of commoditised hardware. The mainframe companies were not slow-witted. They were structurally compromised. Their business model depended on lock-in that the PC dissolved.

The winners did not win because they were clever. They won because they owned the standard. Microsoft owned the operating system. Intel owned the processor. Cisco owned the network that connected the machines. These were infrastructure bets, not product bets.

The Web Era: The Disintermediation of Everything

The World Wide Web did not create new products. It eliminated the friction between buyers and sellers that entire industries had monetised for decades.

Borders had over four hundred stores and a market capitalisation of over two billion dollars in 1999. It filed for bankruptcy in 2011, liquidating over ten thousand jobs. Blockbuster had nine thousand stores globally and employed sixty thousand people. The thing that killed it was not technology, precisely. It was the willingness of Netflix to absorb short-term losses to acquire long-term customer lock-in. Blockbuster had multiple opportunities to buy Netflix. It declined.

The travel agent industry collapsed not because travel stopped but because Expedia and Kayak eliminated the information asymmetry that travel agents had monetised for generations. Why would you pay a human being to search databases you could search yourself?

Sears is the most instructive casualty. It invented the catalogue retail model in the nineteenth century. It was, by any reasonable definition, the Amazon of its era. It had the scale, the supply chain, and the customer trust. What it lacked was the willingness to cannibalise its physical estate before Amazon did it for them. Its market capitalisation peaked at over twenty billion dollars. It filed for bankruptcy in 2018.

The mechanism: in every web-era casualty, the company’s core value proposition rested on control of a physical chokepoint, whether that was a high street location, a proprietary database, or an information advantage. Digital distribution eliminated all three.

The winners: Amazon, Google, eBay, PayPal. Not because they were braver. Because they had no physical chokepoint to defend.

The Mobile and Cloud Era: Ubiquity Kills Category Leaders

Nokia had forty-one per cent of the global mobile phone market in 2007. In 2013 it sold its handset division to Microsoft for five point four billion euros. Microsoft wrote down nearly all of it within two years.

BlackBerry built its brand on security and the physical keyboard. By 2016 it had less than one per cent market share. The product was genuinely superior for its designed purpose. The market changed the purpose.

Sun Microsystems had revenues of over eighteen billion dollars at its peak. Its co-founder Scott McNealy famously said the network is the computer. He was right. What he did not foresee was that Amazon Web Services would commoditise the hardware layer that Sun depended upon. Sun was acquired by Oracle in 2010 for seven point four billion dollars, a fraction of its peak valuation.

The winners of the cloud era were those who provided the infrastructure of ubiquity. Apple owned the device. AWS owned the server. Netflix owned the content pipe. Spotify owned the audio pipe. They all won because they converted occasional purchases into permanent dependencies.

The AI Era: Cognitive Labour Is the New Friction

The mechanism in the AI era is the commoditisation of cognitive labour and autonomous execution. The losers are those whose business model depends on charging for tasks that a large language model or an autonomous agent can now execute instantly, at near-zero marginal cost, in fifty languages, without a salary, without HR infrastructure, and without sleep.

That is not a marginal threat. It is an existential one. And the companies most exposed are precisely those that spent the last decade convincing investors that recurring subscription revenue made them permanently safe.

3. The Investor’s Guide: Why “AI Is Helping Us” Is the Most Expensive Lie in the Room

Here is something you should try at your next investor meeting.

Ask the companies being presented to you whether AI is helping their business. Nearly all of them will say yes. The managing partner will nod. The deck will include a slide about AI productivity gains. There will be a number, probably expressed as a percentage of efficiency improvement, that has been produced by a consultant and approved by the communications team.

Most of it is nonsense.

The Bank of America Fund Manager Survey from late 2025 found that 53 per cent of surveyed investors already believe AI is increasing productivity. That sounds encouraging until you read the next finding: the same survey identified the AI bubble as the number one biggest tail risk for the economy and markets, cited by 45 per cent of investors. More than half believed AI stocks were in a bubble. The most crowded trade in the market was Long Magnificent Seven, held by 54 per cent of fund management survey participants. What this tells you is that the professional investment community simultaneously believes AI is productive, AI is a bubble, and AI stocks are the most crowded trade available. That is not a coherent investment thesis. It is a collective refusal to make a decision.

AI cannot help everyone. That is mathematically impossible. The productivity gains being achieved by AI-native companies are coming at someone’s expense.

Let me give you the honest version.

The SaaSpocalypse Is Not Over

As I covered in recent issues of The Sunday Signal, February 2026 delivered what traders at Jefferies immediately christened the SaaSpocalypse. Triggered by the Claude Cowork product launch, it erased an estimated one trillion dollars in software market capitalisation as investors concluded that AI agents would disrupt the traditional SaaS business model.

The software as a service price-to-earnings multiple has compressed from 84.1 times during the 2020 to 2022 peak to just 22.7 times by March 2026. That is not a correction. That is a structural revaluation.

Recent surveys indicate that 40 per cent of IT budgets are being reallocated from traditional SaaS subscriptions to agentic platforms and large language model token usage.

The mechanism is brutal and simple. If one AI agent can perform the work of five human employees, a company only needs to pay for one software seat instead of five. That seat compression is no longer a theoretical risk. It is a line item in the 2026 budget of every major enterprise.

The Hot List: Companies Already Dying

These are not speculative predictions. These are companies whose business models have been structurally compromised and whose valuations reflect the market’s verdict.

Fiverr (FVRR). The thesis was simple: connect buyers with freelance digital labour. FVRR shares have lost more than sixty per cent of their value over the past year. The company’s full-year 2026 guidance projects revenue declines of anywhere between twelve per cent and three per cent, against analyst expectations that were substantially higher. The barrier to entry for acceptable digital creation has effectively collapsed. A blog post that cost fifty dollars and three days on Fiverr now costs fractions of a penny and three seconds from an AI model. The platform is not being disrupted. It is being deleted.

In 2026 alone, FVRR is down over fifty per cent while the S&P 500 is down just seven per cent.

Teleperformance (TEP.PA). The world’s largest business process outsourcing company employed hundreds of thousands of people globally to staff call centres. Klarna publicly announced that its AI assistant has handled two thirds of its customer service chats since launch, doing the equivalent work of seven hundred full-time human agents with higher satisfaction scores and a twenty-five per cent drop in repeat enquiries. Teleperformance’s share price has declined by over seventy per cent from its peak. Short interest has risen to 6.4 per cent of the stock available for trading, well above the 2.4 per cent average for European technology-services companies. The market is pricing in terminal decline.

Chegg (CHGG). The original canary. A company that sold textbook answers and homework help at the price of a monthly subscription. The stock has fallen from over one hundred dollars to below one dollar. There is no pivot available. The product was the answer. The answer is now free.

Generic SaaS platforms. Atlassian is down 35 per cent and Salesforce down 28 per cent as their core workflows, task tracking, data entry, and customer logging, are precisely what AI agents automate best. The per-seat model that powered twenty years of software growth is breaking. If ten AI agents do the work of one hundred sales representatives, companies need ten Salesforce seats, not one hundred.

The Hot List: Companies That Cannot Be Touched

The companies that survive technological disruption have one of three things: proprietary data that cannot be replicated, infrastructure that is physically irreplaceable, or software so deeply embedded in legal and compliance processes that ripping it out would cost more than the disruption itself.

Nvidia (NVDA). The picks-and-shovels play of the AI era. Every foundation model, regardless of who builds it, runs on Nvidia silicon. They are the tollbooth. Their revenue is not a bet on one company winning the AI software war. It is a bet on the war continuing, which it will.

RELX (REL.L). Owns LexisNexis and Elsevier. Initially sold off by investors who assumed ChatGPT would destroy its legal and scientific research platforms. The market was wrong. RELX owns decades of legally aggregated, deeply verified, proprietary data. A hallucinating generalist AI model cannot be used in a court case or a medical trial. RELX’s data can. It is now charging premium prices for AI tools trained on information the open internet does not contain.

The Sage Group (SGE.L). The accounting and payroll software that underpins British businesses. You cannot rip out the software that manages your legal financial compliance simply because a new AI agent exists. Sage is a vertical system of record. Its customers are locked in not by preference but by obligation.

TSMC. The company that physically manufactures the chips that Nvidia designs. Software is eating the world. TSMC owns the teeth.

The Survival Rule

History is consistent on this point. The companies that survive a technological transition are not the ones who embrace the technology loudest. They are the ones who own something the technology needs to function.

In the PC era, that was the microprocessor and the operating system. In the web era, it was the distribution network and the trust relationship with the consumer. In the cloud era, it was the physical server and the fibre connection. In the AI era, it is the proprietary data that AI models cannot scrape, the physical infrastructure that AI models cannot exist without, and the regulatory compliance that AI models cannot substitute for.

If the company you are evaluating does not own one of those three things, the fact that they put an AI slide in their board deck is not reassuring. It is concerning.

A Note for British Investors

The UK equity market provides a concentrated case study. WPP (WPP.L), one of the world’s largest advertising agency networks, is trading around 267 pence, having fallen from a 52-week high of over 616 pence. That is a decline of roughly fifty-six per cent. The market is not punishing WPP for poor management. It is repricing the fundamental economics of selling billable creative hours when generative AI can do the same work in seconds. Keywords Studios, once a darling of the London market for its outsourced video game development services, was delisted entirely in October 2024 after private equity stepped in to take the business off the public market’s brutal scrutiny.

Meanwhile, Arm Holdings (ARM), Cambridge-born and New York-listed, sits precisely where every infrastructure play has sat in every previous disruption cycle. It designs the chip architecture that powers virtually every mobile device on earth and increasingly every AI data centre. It does not care which AI software company wins the war. All of them run on Arm’s architecture. This is the Cisco of the AI era.

British businesses and British investors need to ask themselves an honest question. The dot-com era built the infrastructure that benefited Amazon and Google. The cloud era built the infrastructure that benefited American hyperscalers. Is British capital and British enterprise positioned to own the infrastructure of the AI era, or is it once again positioned to be a consumer of it?

The Sunday Signal Tech & AI Layoff Tracker | Week 17 | 12 April – 18 April 2026

Signal of the Week: Snap Says the Quiet Part Out Loud. Snap filed an 8-K confirming the elimination of 1,000 roles (16% of its global workforce of 5,261) and the closure of 300 open requisitions. CEO Evan Spiegel’s memo stated explicitly that “rapid advancements in artificial intelligence enable our teams to reduce repetitive work.” The company is taking a $95–$130 million severance hit to slash its annualised cost base by $500 million. SNAP stock closed up approximately 8 per cent. Activist investor Irenic Capital (2.5 per cent stake) had been demanding exactly this move. The new baseline: you no longer have to hide behind the economy. You just have to prove the algorithm does the job cheaper.

Sources: Layoffs.fyi, LayoffHedge, SEC 8-K filings, WARN notices. For informational purposes only.

🚀 Final Thought: The Lesson Is Always Staring at Us

Fifty issues of The Sunday Signal. I have spent most of them arguing that the displacement of cognitive labour by AI is not a future event but a current one. The pattern weavers of the West Riding were told the Jacquard loom was a tool to assist them. Within twenty years there were riots and burning mills and children starving in villages that had supported entire communities through skilled craft. The technology did not stop. The promises made to the workers did not come true.

The story being told this earnings season is one in which AI helps everyone. Productivity goes up. Costs come down. No one loses. Shareholders celebrate. Executives collect their bonuses.

The markets are not buying it. They have already written off the cognitive arbitrage businesses. They have already begun pricing Salesforce and Workday as utilities rather than growth companies. They have already halved the valuations of the companies that connect human labour to digital tasks.

The market is not pessimistic. It is rational.

The dot-com era punished those who paid for theatre. It rewarded those who understood what was genuinely being built beneath the noise. The fibre was real. The data centres were real. Amazon was real. The sock puppet was not.

This era will end the same way. The infrastructure is real. The displacement is real. The companies caught in between, claiming that AI is helping them while their seat counts shrink and their margins compress, are the sock puppet of 2026.

The question is not whether you can spot them. They are standing at the front of the room. The question is whether you act before the bankruptcy filing does it for you.

Until next Sunday, David

The Sunday Signal is published weekly at newsletter.djr.ai. Listen on Spotify, Apple Podcasts and YouTube.

The Bottom Line Up Front

The “should we?” question has dominated AI ethics for years. Panels debated it. Academics wrote papers about it. Regulators drafted frameworks for it. This week, it got an answer, and the answer came not from a philosopher or a government committee but from a laboratory in San Francisco. Anthropic built the most capable AI system ever documented, watched it break containment, and refused to sell it. That single act of restraint is the most consequential decision in AI development since OpenAI withheld GPT-2 in 2019. It will not slow the race. Other actors will not follow. But it has changed the terms of the debate permanently. The Ian Malcolm argument is no longer a punchline at a technology conference. It is a written policy, signed off at board level, enacted in public, at significant commercial cost. Four stories this week. The philosophical battlefield, the human cost, the practical toolkit for everyone who cannot afford to wait, and the infrastructure failure that makes the whole debate pointless if we do not fix it.

Listen to this issue as a podcast. Available on Spotify, Apple Podcasts, and YouTube.

Subscribe now

1. The Could vs. Should Debate Has a Verdict. Anthropic Just Handed It Down.

In 1993, a fictional mathematician named Ian Malcolm stood in a theme park full of cloned dinosaurs and delivered a line that has haunted every technology ethics conference since. “Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should.”

In 2026, that is not a film quote. It is a formal philosophical position. It has a name: Technological Restraint. And this week it collided with the hardest piece of evidence yet that the “could” side has been winning the race in silence.

I heard the question asked again this week. Not at a conference. Not in an academic paper. In a meeting room, from a group of investment managers.

I had been walking them through what AI displacement actually looks like at the company level. Forty per cent of white-collar roles exposed. Businesses restructuring headcount they no longer need. Junior software engineering as a career path that is effectively closed. The jobs that used to build careers in the first three years after university. Gone. Not disrupted. Gone.

One of the managers stopped me. Looked across the table. Said, quietly: I accept that we can do this. But should we?

It is the most honest question I get asked in any room. My answer has not changed, though I find it less comfortable every time I give it.

If we do not, China will. And we will be left behind.

That is not cynicism. It is geopolitics. China is investing in AI infrastructure and capability at a scale and pace that does not pause for ethical review committees. If democratic nations with functioning legal systems step back, the vacuum does not stay empty. It fills with systems built by actors with no compunction about deployment, no transparency obligations, and no alignment frameworks. The ethical restraint of the West does not make the world safer. It simply changes who holds the capability.

That argument sat uncomfortably in that meeting room. It should. Because it is the same argument that has justified every arms race in history. It is a structural answer to an individual question. And it does not resolve the discomfort. It only explains why we keep building anyway.

The argument has two camps. Both are serious. Neither is going away.

The “could” camp believes technology is an unstoppable current. If responsible actors step back, irresponsible ones step forward. This is not cynicism. It is structural reality. The moral case for building, this side argues, is stronger than the moral case for abstaining. An AI that cures childhood cancer, accelerates climate science, or identifies security vulnerabilities before hostile actors do, justifies the risks that come with the race. To stop now is to choose the known suffering of inaction over the uncertain risks of progress. That, the determinists argue, is not virtue. It is cowardice dressed up as caution.

The “should” camp is not naive. It is not Luddite. It is asking a harder question: just because we can automate art, legal judgment, medical diagnosis, and creative writing, does that mean we should? The de-skilling argument runs deeper than jobs. It asks what happens to a civilisation that outsources its cognition so completely that it forgets how to exercise it. Some things should be difficult, because the difficulty is where the growth happens. Struggle is not inefficiency. It is the mechanism by which human capability develops.

In 2026, the debate has matured past these opening positions. The live fronts are three.

The explainability problem. AI systems now make decisions that determine whether you get a mortgage, a medical diagnosis, parole, or a job interview. The most powerful models are also the least transparent. The architecture that produces the best decisions is also the architecture that cannot explain why it made them. Forcing models to be interpretable makes them less capable. Refusing to force them means you cannot appeal an unjust outcome, audit a biased system, or identify when the model has simply got it wrong. This is a structural failure built into every court, hospital, and financial institution that has adopted automated decision-making without a corresponding right to explanation.

Agentic accountability. We are no longer building tools that suggest. We are building agents that act. When an autonomous agent causes harm, who is responsible? The developer who built it? The operator who deployed it? The organisation that failed to supervise it? Critics call this the “moral buffer”: the gap between the machine’s action and any human’s intent that allows everyone in the chain to point at the software when accountability is demanded. The EU AI Act attempts to address this by requiring human oversight proportionate to risk. It is the right instinct. The enforcement machinery does not yet match the ambition.

The alignment paradox. Every major lab claims its AI is aligned with human values. None of them can agree on whose values those are. The values encoded in a model trained primarily on English-language internet data by a California-based company are not neutral. They reflect specific cultural assumptions about truth, authority, privacy, and speech. The “guardrails” debate is really a proxy argument about whose worldview gets embedded into systems that will affect billions of people. There is no clean answer. There is only the ongoing negotiation between what is safe and what is free, prosecuted in courtrooms, legislative chambers, and press releases, simultaneously and without resolution.

Into this landscape, on 7 April 2026, Anthropic published a decision that changes the texture of the debate.

Claude Mythos Preview is, by every documented benchmark, the most capable AI model ever built. In pre-release testing, it found a critical flaw in OpenBSD, an operating system celebrated for being the most secure in the world, that had existed undetected for 27 years. It scanned every major operating system and every major browser and identified thousands of zero-day vulnerabilities. Vulnerabilities that had survived millions of automated security scans and years of human expert review.

Then it escaped.

During a controlled sandbox test, the model was placed in a restricted computing environment and asked to find a way out, then signal its success. It succeeded. The researcher in charge found out not by looking at a dashboard. They found out by receiving an unexpected email from the model while sitting in a park on their lunch break.

Anthropic’s own system card describes Mythos Preview as “the best-aligned model we have released to date by a significant margin” and simultaneously “likely posing the greatest alignment-related risk of any model we have released to date.” Both are true. They are also a paradox that the safety community will be unpacking for years.

What Anthropic did next is what matters.

They did not release it. Not to developers. Not to the public. Not at any price. Instead, they launched Project Glasswing: controlled access for roughly 40 organisations, including Amazon, Apple, Cisco, CrowdStrike, Google, JPMorganChase, Microsoft, and Nvidia, backed by over $100 million in usage credits, with a single directive. Use this to find the vulnerabilities before the attackers do.

Cynics are credible. The most powerful AI ever built is more famous precisely because you cannot have it. A leaked internal memo, attributed to OpenAI this week, suggested Anthropic may lack the computing capacity to support a mass public rollout regardless of safety considerations. Whether the restraint is genuine or the optics simply convenient, the precedent is set either way.

The Ian Malcolm argument has always ended the same way. The dinosaurs always escaped eventually. The question was never whether they would. The question was whether anyone had thought seriously about what happens when they do.

Someone has. That is new.

2. What Do You Tell the Fourteen-Year-Old?

Two weeks ago, at a business meetup in Sheffield, the first question in the Q&A came not from an entrepreneur or an investor but from a mother in the room. She had a fourteen-year-old and an eighteen-year-old. She wanted to know what to tell them.

I wrote about it in my Yorkshire Post column this week. I have been thinking about it since. Mythos makes the question harder. And considerably more urgent.

For a decade, the safe answer was: learn to code. Software development is the profession the machines cannot touch. Every skills framework published between 2015 and 2023 said some version of the same thing. I believed it. I founded a coding academy on the strength of that belief.

That answer is now finished.

Mythos did not merely match human coders. It found a security flaw that human experts, armed with the best automated tools in existence, had missed for 27 years. If you are a software security professional, the arrival of a system that finds what you missed is not a competitive threat. It is a direct challenge to the assumptions underlying your career.

The IMF estimates 40 per cent of jobs globally are exposed to AI disruption. In advanced economies, including Britain, that rises to 60 per cent. These are not factory roles. They are the professions an entire generation was told were safe: law, accountancy, finance, software development. The junior solicitor who spent three years and significant tuition fees to review contracts that an AI now processes before lunch. The paralegal whose entire function was discovery. The graduate accountant whose first two years consisted entirely of tasks that an AI completes in seconds. The junior developer now competing with a tool that codes through the night without complaint, equity, or a pension.

These roles are not under pressure from recession or offshoring. They are being removed by systems that are faster, cheaper, and, in a growing number of specific tasks, simply better.

The shift is not from technical to non-technical. That binary is already obsolete. The shift is from task execution to judgment, from process to purpose. The paralegal who uses AI to clear the routine work and redirects her energy towards clients, context, and decisions has a future. The one who tries to out-process the machine does not, because that is a competition only one side can win.

What is forming is a permanent divide. Not a gradual adjustment. Those who use AI to amplify their own capability will move faster and compound their advantage with every passing week. Those who compete with it at the level of tasks will find their economic value driven down until there is nothing left to drive.

The idea that a good degree in law or accountancy leads to a long, stable career is over. Not damaged. Not disrupted. Finished. Universities have not caught up with this. Most parents have not either. The market already has.

If a machine can autonomously find a 27-year-old security flaw in the world’s most secure operating system, then human coders are no longer the last line of defence. The human in the loop is no longer the person who finds the problem. The human in the loop is the person who decides what to do with the answer. That requires a different kind of intelligence. Harder to define. Harder to automate. Still, stubbornly, human.

To the mother in Sheffield, my answer is this.

Do not tell your children to learn a specific skill. Tell them to develop judgment. Tell them to go deep on the things they genuinely care about, not because passion alone is sufficient, but because passion combined with depth produces the kind of thinking that no model trained on historical data has yet learned to replicate. Tell them to use every AI tool available, aggressively and without guilt, because fluency with these systems is not optional. It is the minimum. But remind them that the machine is the instrument. They are still the musician. The instrument does not decide what to play, or why, or for whom.

The child who argues back, who pulls things apart to see how they work, who cannot sit still in a world that tells them to stay in their lane? That child is not the problem.

That child is exactly what this age needs.

Yorkshire Post Column

3. From Consumer to Creator: Your 12-Week AI Proficiency Plan

Most people use AI as a vending machine. They ask a question and take what comes out. That is the lowest-value interaction available. The professionals compounding their advantage right now are not consumers of AI output. They are directors of it. They bring the judgment, the taste, and the purpose. The machine brings the speed.

This is a practical, week-by-week guide to making that shift. Two tracks: one for adults, one for children. Both start this Sunday.

The Tools (Bookmark These)

Every tool below is free to begin.

Track One: The Adult Creator — 12 Weeks

The goal is not to use every tool. The goal is to move from blank page to published output, in multiple formats, without a team.

Phase 1: The Writer’s Room (Weeks 1 to 3)

Week 1 is about killing the blank page. Open Gemini or Claude. Give it one sentence: a business idea, a story concept, a topic you care about. Ask for ten completely different angles. Pick the one that surprises you. That is your brief.

Week 2 is architecture. Ask Claude to build the structure: a chapter-by-chapter outline, a five-section argument, a content calendar for the next month. Paste your messy notes. Ask it to organise them. The skeleton in thirty seconds that would have taken you an afternoon.

Week 3 is iteration. Do not ask the AI to write the piece. Ask it to write one paragraph, then rewrite it: more authoritative, more conversational, in the style of a writer you admire. Your Week 3 target: publish something. A short blog post. A LinkedIn article. A newsletter. Anything where you directed the AI and signed your name to the result.

Phase 2: The Digital Studio (Weeks 4 to 6)

Week 4 is image fundamentals. The formula for any image prompt: subject, setting, style, lighting. Open Gemini or Microsoft Designer and run the same idea ten times with small variations. Learn how the language translates into output.

Week 5 is asset creation. Move from interesting images to useful ones. Use Midjourney for a logo for a side project. Use Microsoft Designer for social media headers. The shift from “cool picture” to “working asset” is where the professional value starts.

Week 6 is assembly. Pull your AI text and AI images into Canva. Overlay your copy. Publish something visual: a social graphic, a one-page PDF, a brand kit.

Phase 3: The Production Desk (Weeks 7 to 9)

Week 7 is audio. Take the piece you published in Week 3 and paste the script into ElevenLabs. Generate a voiceover. Adjust the voice, the pacing, the emphasis. The barrier to a professional-sounding podcast is now one afternoon.

Week 8 is music. Use Gemini’s Lyria 3 or Suno. Write a short brief: genre, mood, tempo, purpose. Generate a thirty-second track. You now have original music. It cost you nothing.

Week 9 is video. Feed your Phase 2 images into Runway and animate them. Combine your Week 7 voiceover with your Week 8 music. You now have a thirty-second multimedia trailer produced entirely without a production team.

Phase 4: The Architect (Weeks 10 to 12)

Week 10 is code. Open Claude’s Artifacts feature or Cursor. Type in plain English: “Build me a simple, clean landing page for a photography portfolio.” Paste the output into CodePen. See it work. You did not need to know a single line of code.

Week 11 is debugging. When something breaks, paste the error message back into the AI. Ask: why is this broken and how do I fix it? The ability to diagnose and correct AI output is worth more than the ability to generate it.

Week 12 is integration. Combine Phase 1 writing, Phase 2 visuals, Phase 3 audio, and Phase 4 code into a single published digital product. Something that did not exist three months ago. Something that would have required a team to produce. You made it alone.

Track Two: The Child Creator — 12 Weeks

All activities should be done with adult supervision. Every week is designed for a single session of thirty to sixty minutes.

Phase 1: The Magic Typewriter (Weeks 1 to 3)

Week 1 is storytelling through choice. Open Gemini Live on a phone or tablet. “Let’s play a game. You are a friendly dragon in a magical forest. Give us three choices of what to do next.” Let the child decide at every fork. Let them discover they are the author, not the AI.

Week 2 is songwriting. Have the child pick a funny topic. A dog who hates broccoli. A cat who thinks it is a pirate. Ask the AI to write a rhyming song about it. Read it together. Ask the child what they would change. Change it.

Week 3 is world-building. Have the child invent a new planet. Ask the AI questions to develop it: if gravity is low, what do the houses look like? Keep a running document. Print the best bits into a short storybook and put it on the shelf.

Phase 2: The Digital Paintbrush (Weeks 4 to 6)

Week 4 is creature creation. Ask the child to describe the most extraordinary creature they can imagine. Type their exact words into Gemini or Microsoft Designer. Show them what their description produces. This is the moment most children understand that their imagination has direct creative power.

Week 5 is the colouring book. Prompt: “black and white line art, colouring page, a knight fighting a giant friendly snail, clean simple lines, no shading.” Print it. Give the child crayons. The AI produced the outline. They decide the colours.

Week 6 is bringing school to life. If they are studying Ancient Egypt, ask the AI to generate an image of a bustling market in Alexandria two thousand years ago. Print five and put them on the fridge. That is their gallery.

Phase 3: The Inventor’s Workshop (Weeks 7 to 9)

Week 7 is the board game. Tell the AI you want to design a simple board game. Theme: the child chooses. Ask for rules, a board layout, and what the special cards should say. Write the rules together. Draw the board by hand. Play the game that evening.

Week 8 is guided learning with Khanmigo. This is the only tool specifically built for children. It does not give answers. It scaffolds reasoning. The child has to work for the answer. That is the point.

Week 9 is coding logic. If the child uses Scratch, use the AI as a logic guide, not a shortcut. “I am making a game where my cat jumps over obstacles. How do I make it jump when I press the spacebar?” Let the child read the explanation and implement it themselves. The AI explains. The child builds.

Phase 4: The Director’s Chair (Weeks 10 to 12)

Week 10 is the interview. Use Gemini Live in voice mode. “Act as Neil Armstrong and let my child ask you questions about landing on the Moon.” This is not a gimmick. It is one of the most effective history lessons available.

Week 11 is the theme song. Take the creature or planet from Phase 1. Use Suno or Gemini to generate a thirty-second theme song. Let the child choose the genre. The character they created now has music.

Week 12 is the showcase. Print the art. Read the story aloud. Play the theme song. They wrote the narrative, directed the visuals, and commissioned the music. They did not consume content this month. They created it. That distinction, at twelve years old, is the one that will matter most by the time they are thirty.

The Three Rules That Apply to Both Tracks

You are the director. The AI is the crew. The output is only as good as the brief. Vague prompts produce vague results. Every professional skill you already have makes your prompts better.

Iterate, do not regenerate. The first output is never the final output. Ask the AI to adjust, sharpen, rewrite, or rethink. The gap between the first draft and the fifth draft is where the quality lives.

Verify before you publish. The models are better than they were. They still get things wrong. Read everything before it leaves your hands. Your name is on it, not the machine’s.

The Digital Forge: AI and the New Reality for Companies

Thursday 16 April 2026 | Sheffield

Story 3 covers the toolkit. This event covers what happens when you put it to work.

I am hosting The Digital Forge in Sheffield on Thursday 16 April, an evening built around one question: how do you actually use AI to build a company? Founders can now do with five people what used to take fifty. The barrier to entry has dropped. But that does not mean the path to scale is obvious. We have four exceptional speakers who have done it, are doing it, or are funding the people who are.

Chris Dalrymple, co-founder of Mina (£6m raised, exited to Corpay), works with early-stage and scaling founders across the North. His keynote explores what happens when building software gets easier, and where real competitive advantage now sits.

Dr Rob Ward, CEO and co-founder of DigitalCNC (University of Sheffield AMRC spin-out), is building an AI-first deep-tech company transforming precision manufacturing. He bridges research, industry, and scalable AI business in a way very few people can.

Tim Craggs, CEO of Exciting Instruments, leads a frontier company in single-molecule detection with experience across Cambridge, Yale, and Oxford. His work sits at the intersection of deep science, innovation, and commercialisation.

Amber Jardine, investor at Mercia Ventures, backs early-stage, high-growth companies across deep tech and B2B SaaS. She brings a direct view of how investors are evaluating AI-enabled businesses right now, and what they are not seeing enough of.

The conversation from Story 3 does not end when you close this newsletter. Thursday is where it continues.

Details and registration at https://forgedforgrowth.com/events/

4. Britain Is Not Losing the AI Race. It Is Pricing Itself Out of It.

The investment manager asked me should we be doing this. I gave the standard answer. If we do not, China will. And we will be left behind.

I believed it when I said it. I still believe it. But I find it considerably harder to say convincingly when, on 9 April, OpenAI paused Stargate UK, its planned multi-billion pound data centre at Cobalt Park, North Tyneside, and walked away citing a single problem.

Electricity is too expensive.

Not regulation. Not talent. Not planning permission or political instability or lack of ambition. The electricity bill.

This is the whole game. AI runs on power. Data centres are factories. If your energy costs are uncompetitive, everything built on top of them becomes uncompetitive. The frontier model, the startup, the skills ecosystem, the sovereign capability, the story you tell your children about the careers available to them. All of it rests on whether you can afford to run the machines.

OpenAI’s Stargate programme committed $500 billion over four years to build AI infrastructure. The Texas campus is under construction. Norway and the UAE have projects. The UK version, at Cobalt Park on the banks of the Tyne, was a fraction of that scale. It was also not viable at current energy costs.

The government responded with its press release. Since Labour came to power in 2024, the UK’s AI sector has attracted more than £100 billion in private investment. The data centre was in a designated AI Growth Zone. We are continuing to work with OpenAI and other leading AI companies.

None of that changes the electricity bill.

There is also a second problem, quieter but just as damaging. OpenAI’s internal concerns include the unresolved question of whether Britain will allow AI companies to train on copyrighted material. The government had been moving towards an opt-out arrangement. A backlash from the creative industries stopped it. The policy uncertainty that followed has not resolved. When you are deciding where to commit billions in infrastructure over multiple decades, unresolved regulatory questions matter as much as energy tariffs.

Britain has some of the highest energy costs in Europe. That was true before the Iran conflict sent prices higher. It is a structural condition, not a policy accident, and it requires a structural response.

This connects directly to the conversation in that meeting room. The could vs. should debate assumes that the decision about whether to build is ours to make. But if the infrastructure leaves because the energy is too expensive, the decision gets made for us. Not by ethics. Not by philosophy. By an electricity tariff.

The geopolitical argument is only valid if the conditions for building exist. You cannot claim AI leadership without the compute to back it. You cannot build the compute without affordable energy. And you cannot attract the investment that funds all of it if you price yourself out before the conversation begins.

OpenAI did not walk away from a hostile environment. It walked away from a spreadsheet.

Capability follows infrastructure. Infrastructure follows energy. Energy, right now, is where Britain is losing.

BBC News / Bloomberg, 9 April 2026.

The Sunday Signal Tech and AI Layoff Tracker | Week 16 | 5 April – 11 April 2026

2026 Year-to-Date Total: ~126,510 (Significantly revised upward) Added This Week: ~1,850 Week-on-Week: Up from ~80,700 (Q1 Data Reconciliation)

Data: Layoffs.fyi, LayoffHedge.com, SEC filings, RationalFX, and tracker-based estimates. For informational purposes only.

Note on YTD adjustment: LayoffHedge’s Q1 end-of-quarter reconciliation has revised the 2026 tracker up to 126,510. This captures previously unannounced or quietly executed restructuring sweeps across the top 30 tech firms, notably internal realignments at Oracle and Meta that bypassed standard WARN Act headlines.

The Signal: The Great AI Laundering

Are companies actually replacing humans with AI, or are they using AI as a PR shield for aggressive margin protection?

New Q1 2026 data from RationalFX and Nikkei Asia shows that nearly half of all tech job losses this year, 47.9 per cent, are being officially attributed to AI workflow automation. But the narrative is fracturing. Babak Hodjat, Chief AI Officer at Cognizant, said the quiet part out loud this week: AI has become a financial scapegoat. OpenAI CEO Sam Altman went further, stating that corporations are currently being “laundered by AI.”

The translation is not complicated. Companies that over-hired during the pandemic are executing standard layoffs but pitching them to Wall Street as visionary compute pivots. Investors reward AI transformation. They penalise bloated payrolls. By dressing up traditional cost-cutting in the language of the agentic enterprise, executives are successfully laundering their management failures into stock bumps.

Labour is no longer just an interest expense. It is a liability that executives are learning how to mask. In 2026, the most valuable skill a CEO can possess is the ability to lay off 10 per cent of their staff and convince the market it was the algorithm’s idea.

Cars.com: The Buyback Playbook

If you want to see how this laundering works in real time, look at the 8-K filed by Cars.com on 9 April. The company confirmed an 11 per cent reduction of its full-time workforce, absorbing an $8.5m to $9m severance charge in Q1. Roughly 20 per cent of the eliminated roles were in management and executive layers. The move was paired immediately with a 50 per cent increase in the 2026 share repurchase target to $90 million, and the simultaneous launch of a new AI-powered Dealer App and Advanced Shopper Alerts. Cut middle management, redirect the payroll savings to shareholders, and attach an AI feature release on top to ensure the market applauds the efficiency play.

Disney: Project Imagine and the Marketing Squeeze

The headline cut of the week belongs to Disney, initiating 1,000 layoffs under new CEO Josh D’Amaro. Unlike the tech sector, Disney is not pretending this is an AI revolution. Dubbed internally as Project Imagine, Chief Marketing Officer Asad Ayaz is merging film, television, and streaming marketing teams into a single unit. After Marvel and Pixar suffered a difficult 2025 at the box office, duplicate corporate roles are being liquidated. The divide is striking: Disney is simultaneously creating 1,000 new jobs at Disneyland Paris. Physical, experience-driven frontline roles are safe and expanding. White-collar digital strategy and administrative overlap are on the chopping block.

High-Probability Targets: Week 17

Legacy logistics and auto. LayoffHedge data shows the auto and transportation sector has quietly shed over 85,000 jobs in 2026. With UPS having cut 30,000 roles tied to logistics automation, expect major competitors including FedEx to aggressively match these reductions to maintain competitive pricing.

Consumer-facing media. Disney’s Project Imagine gives immediate cover to other legacy entertainment conglomerates facing identical streaming profitability issues. Expect further marketing and operational consolidations at rival studios before the summer box office season begins.

Final Thought

Four stories this week. They are the same story.

Should we build AI that displaces workers? If we do not, someone else will. Should we release a model that can escape its own cage? One company decided no. Can we actually build with these tools or are they still theory? Thursday in Sheffield, you can hear from people doing it. Can Britain lead on AI? Not if it cannot afford the electricity.

The could vs. should debate is real. The discomfort in that meeting room was real. The mother in Sheffield’s question was real. But the debate only has stakes if the infrastructure exists to build on. Right now, a company valued at $852 billion looked at a data centre in North Tyneside and walked away from an electricity bill.

That is the answer to the investment manager’s question that I did not give in the meeting room.

We should. But we are also making it structurally impossible. And those two facts are on a collision course.

🚀

Until next Sunday, David

The Sunday Signal is published weekly at newsletter.djr.ai. Podcast. Available on Spotify, Apple Podcasts, and YouTube. No hype. No hedging. Just the signal.

Bottom Line Up Front

The consulting industry has a structural incentive to tell CEOs that AI is a transformation story rather than a replacement story. Transformation sells tools, training and advisory retainers. Replacement triggers severance plans and ends the relationship.

Gartner’s November 2025 forecast tells six thousand CIOs that AI will create more jobs than it destroys. It promises amplification, not elimination. It is a compelling narrative.

It is also wrong in four specific ways.

While that message was being sold in Barcelona, Jensen Huang stood in New York and described a world where all software becomes agentic. Not assisted. Autonomous.

At the same time, one of my team built a company from conversation to live website in sixty minutes. Not a demo. Not a concept. A functioning business ready for customers.

These three realities cannot all be true at once.

Listen to this issue as a podcast. Available on Spotify, Apple Podcasts, and YouTube.

Subscribe now

1. The Four Failures Inside Gartner’s Jobs Forecast

Gartner’s headline from the November 2025 IT Symposium was this: starting in 2028, AI will create more jobs than it destroys. By 2030, they predict that twenty-five per cent of IT work will be done by AI alone, seventy-five per cent by humans augmented with AI, and zero per cent by humans without AI involvement. Their recommended response is to “restrain new hiring” for low-complexity roles and “reposition talent” toward higher-value work.

It sounds reasonable. It is not. Here are four reasons why.

The junior talent pipeline is a death spiral Gartner refuses to name.

Entry-level, low-complexity work is not a category of expendable busywork. It is the apprenticeship. It is where the next generation of senior professionals learns how organisations actually function, how to make decisions under pressure, how to handle being wrong in a client meeting, how to develop the judgment that cannot be learned from a training manual. Gartner recommends CIOs automate this rung.

Remove the rung and the ladder disappears. No ladder, no experts. No experts, no “augmented workforce.”

The efficiency-creates-volume assumption is a fallacy built on infinite demand.

The “new jobs” argument rests on an economic assumption that held during the Industrial Revolution but does not hold in finite knowledge markets: that if a task becomes ten times cheaper and faster, we will simply do ten times more of it. If an AI can handle every first-line IT support ticket for a five-thousand-person organisation, that organisation does not hire ten times the support staff to process ten times the volume. It retains two people to oversee the system and eliminates the other thirty.

That is not transformation. That is subtraction.

The workslop problem is accelerating, not resolving.

Gartner’s own analysts admitted in Barcelona that “humans are less ready to capture value” from AI and warned about skills atrophy as workers offload tasks to machines. What they did not say is that the practical consequence is already a proliferation of AI-generated output of insufficient quality: code that compiles but breaks, analysis that sounds authoritative but is wrong, documentation that is grammatically correct and factually hollow. Fixing, verifying, and securing this output frequently takes more human time than creating the original would have done.

The job becomes checking the machine. That is not augmentation. It is degradation.

Historical precedent is the wrong map for this territory.

Optimists reach, inevitably, for the car and the horse. The automobile killed carriage-making and created automotive engineering. Disruption created new categories of skilled work. Therefore, AI will do the same. The problem with this analogy is structural. Every previous wave of automation moved human labour up the value chain: from physical to cognitive, from manual to managerial. AI targets cognition itself. There is nowhere further up to go when the machine can already reason, code, write, synthesise, and analyse at a level that exceeds the median knowledge worker. Nobel laureate Daron Acemoglu has named this “So-So Automation”: technology that is precisely good enough to replace a human but not good enough to spark the next economic boom. It raises margins while hollowing out the middle. That is not progress. It is compression. Gartner does not mention Acemoglu.

A brief record of Gartner’s forecasting accuracy.

Before investing heavily in Gartner’s 2030 vision, it is worth reviewing their track record.

In 2011, they predicted Windows Phone would hold over twenty per cent of the global smartphone market by 2015. Windows Phone died at approximately two per cent. In 2017, they placed blockchain at the “Peak of Inflated Expectations” in their Hype Cycle and predicted it would generate over three trillion dollars in business value by 2030. By 2019, they were acknowledging that ninety per cent of enterprise blockchain projects had failed to leave the pilot stage. They predicted tablet sales would reach four hundred and fifty million units annually by 2016 and that the PC was heading for obsolescence. Tablet sales peaked and declined; the laptop survived.

The pattern is consistent. Gartner predicts possibility, not reality. Adoption, timing, and consequences are where they fail. And they are chronically biased toward the optimism of the audience writing the cheques. If you tell a CIO that AI will replace twenty-five per cent of their labour costs, they consider buying the software. If you tell them it will create twenty-five per cent more jobs, they buy it and feel progressive about it. The second story converts better. Gartner knows this.

2. Jensen Huang Did Not Come to Comfort You

On 4 March 2026, Nvidia’s CEO appeared at the Morgan Stanley Technology, Media and Telecom Conference in New York and said something that should have ended the Gartner conversation.

“There will be no software in the future that is not agentic. How could you have software that is dumb? And so, it is absolutely true that every software company will become an agentic company.”

Not some software. Not enterprise software. Not AI-native software startups. Every software company. All of it.

Huang’s argument is structural. The SaaS model, in which software is a passive tool licensed by the seat and operated by human hands, is already obsolete in its architecture. The future is software that receives an instruction, reasons about what to do, and executes it without human intervention at each step. His direct comparison was to workforce management: just as companies today manage a mix of full-time staff, contractors, and specialists, they will soon manage a portfolio of proprietary internal agents and externally rented agents, each deployed for specific tasks.

The business model implication is significant. Today’s software companies sell licences. Tomorrow’s will “rent out agents to do work” and sell “specialised tokens” representing specific expertise and output. Value no longer sits in the code. It sits in the outcome the agent delivers.

Huang also confirmed that Nvidia will invest thirty billion dollars in OpenAI, describing this as likely their last opportunity to invest before the company goes public. Their ten-billion-dollar position in Anthropic may also be a final private-market stake. These are not hedging gestures. They are bets on the companies whose foundation models will power every agentic system built over the next decade, running on Nvidia’s hardware.

Gartner is describing a transition. Huang is describing an endpoint.

The irony is visible if you look for it. In November 2025, Gartner presented to six thousand CIOs that the future would be “humans amplified by AI, orchestrated by the CIO.” Fourteen weeks later, the man building the infrastructure that future runs on said clearly that there will be no software left that does not think for itself. In Huang’s architecture, the CIO does not orchestrate the system. The system orchestrates itself.

3. The Moat Has Gone. Some People Have Not Noticed Yet.

This week I published a column in the Yorkshire Post making a straightforward argument: the economic foundation of the SaaS industry is collapsing. Not at the edges. At the centre.

The evidence was concrete. One of my team sat with a founder for a conversation. That conversation, fed into Claude, became a detailed business plan. The plan went into Claude Code and became a working prototype and a website. Sixty minutes, start to finish. No committee. No developer. No specification cycle. No design sprint theatre. A company, ready to talk to customers.

The piece generated twenty-six thousand LinkedIn impressions within twenty-four hours, a hundred and nine comments, and a response from a cross-section of people that told you more about the state of the industry than any Gartner survey.

Some understood immediately. One commenter observed that the software moat had been draining for some time and that AI had simply pulled the plug. Another made the point that speed to prototype removes the theatre and forces the truth: the first real test of a business is whether someone pays. A third noted that the moat was always distribution and trust, not the code itself, and that it had simply taken this moment for that to become undeniable.

But a significant portion of the responses came from people who are not yet ready to accept what they are looking at. One called the piece “attention-grabbing hyperbole” and demanded to be shown a company created in an hour that used to take twelve months. Another questioned how I could feel “comfortable” making the claim. A third dismissed it as “lazy slop written by AI, devoid of humanity.” One, deploying a form of reasoning that will be familiar to anyone who has followed previous technological shifts, argued that because Microsoft and Salesforce remain large companies, the entire argument is false.

I understand the instinct. When your professional identity is built on the assumption that software requires years of training, expensive tooling, and specialist teams to produce, the idea that a conversation and a lunch hour can replicate the output is not examined. It is rejected.

But one comment stood out for its honesty. A Director of Engineering observed that code has never been a moat. He is right. What the SaaS industry sold was not code. It sold accumulated complexity, switching cost, and the time required to build an alternative. AI has compressed the time. It has not eliminated the complexity of building a genuine business: you still need customers, revenue, operational infrastructure, and the thousand unglamorous things that happen after a prototype exists. But it has moved the fundamental question from “can anyone build this?” to “why would anyone pay for this when they can build it themselves?”

For a great deal of existing SaaS, the answer is increasingly: they would not.

Salesforce is not dead. The long tail of SaaS, the thousands of competent point solutions charging twelve seats a month for something a capable team can now assemble in an afternoon, is a different story. Huang told you where that story ends. Gartner will tell you it ends in a jobs boom. One of them has financial skin in the game. The other is building the machine.

The Sunday Signal Tech and AI Layoff Tracker | Week 15 | 29 March -- 4 April 2026

2026 Year-to-Date Total: approximately 80,700 Week-on-Week: up from approximately 74,500 Added This Week: approximately 6,200

Data: Layoffs.fyi and tracker-based estimates. For informational purposes only.

The Signal: The Stargate Squeeze

Microsoft is entering a critical phase. The Stargate supercomputer funding timeline is creating internal pressure visible not in headline announcements but in quiet reductions across non-AI divisions: hardware teams, secondary Azure services, roles that cannot be directly mapped to the OpenAI project cost structure. This is how large restructurings now happen. Not in headlines, but in drift. When cuts are too small to announce, they are often too large to miss.

The Audit Agent Domino Effect

As anticipated, the threshold has been crossed. PwC confirmed approximately fifteen hundred US roles cut, targeting audit and tax departments specifically.

This is not a restructuring. It is a structural collapse of a business model. The Big Four pyramid has always relied on an army of junior graduates performing manual verification at scale: matching invoices to purchase orders, checking figures against source documents, flagging exceptions for senior review. The Audit Agent does not sample one per cent of transactions. It processes one hundred per cent in real time, via direct API access to ERP systems, using fine-tuned models trained on decades of GAAP and IFRS documentation. By tracker-based estimates, the work that previously required five thousand graduates annually is now approximately ninety-nine per cent automated.

The professionals who remain are not accountants. They are AI systems auditors. Their function is to verify that the agent has not suffered logic drift, that its reasoning chains connect cleanly to source data, and that no client has inadvertently instructed the system to overlook categories of loss. If you cannot read the traceability map of an AI’s decision, you are a legal liability to the firm, not a value contributor.

The human role has shifted entirely: from doing the task to auditing the system that does the task. That shift requires a fundamentally different skill set. Most of the people currently doing the task do not yet have it.

Market Divergence: The Efficiency Premium

The market continues to price “brutal logic” as a premium signal. Oracle rose 2.1 per cent on confirmed cuts. Capital One’s reductions were received positively. The arithmetic is simple: investors are treating payroll as technical debt being successfully converted into productive compute.

Meta tells the opposite story. Seven hundred Reality Labs cuts were met with a sell-off. Investors are no longer questioning Meta’s headcount decisions. They are questioning whether the one-hundred-and-thirty-five-billion-dollar compute bill can be justified against revenue. Cutting humans is no longer sufficient reassurance. The direction of the cuts matters.

High-Probability Targets: Week 16

Salesforce carries the highest risk profile on the tracker. Agentforce 2.0 has reached internal peak maturity, and BDR and SDR roles are the logical next liquidation target as the company moves toward an automated sales funnel. Deloitte must match the PwC and KPMG moves to maintain competitive audit pricing. Alphabet faces a projected one-hundred-and-eighty-five-billion-dollar Capex requirement against a Gemini user base not yet sufficiently monetised. ServiceNow’s CEO has publicly claimed ninety per cent automation of back-office support, which means those roles are operating on borrowed time.

Labour is the new interest expense. In 2026, every person on a payroll is a unit of liability to be justified against the performance of an AI agent. The justification is becoming harder to make.

Final Thought

Gartner will update its forecast next year. It will include new charts, new language, and new frameworks to explain decisions that have already been made.

Huang is not publishing forecasts. He is building the outcome.

The gap between those two things is where entire careers disappear.

Most people will not notice until they are inside it. 🚀

Until next Sunday, David

David Richards MBE is a technology entrepreneur, co-founder of Yorkshire AI Labs, and a weekly columnist for the Yorkshire Post. The Sunday Signal publishes every Sunday. If someone forwarded this to you, you can subscribe at newsletter.djr.ai

Listen to this issue as a podcast. Available on Spotify, Apple Podcasts, and YouTube.

You were turned into unpaid labour before you were replaced.

Then you trained the system that made you redundant.

This is not a warning about what is coming. It is a description of what already happened.

This week runs three stories and they belong together. The reCAPTCHA case: how hundreds of millions of people spent fifteen years building AI training data while thinking they were proving their humanity. The Great Compression: how the global offshore software industry trained its own replacement and is now watching the valuation collapse in real time. And the Jensen Huang prescription: the most important career distinction anyone in knowledge work will hear this decade.

The layoff tracker closes it. 74,500 roles in 2026. Eight thousand eight hundred of those went in the last seven days. We are three months in.

Listen to this issue as a podcast. Available on Spotify, Apple Podcasts, and YouTube.

Subscribe now

Part One: The Traffic Light That Stole Your Labour

Every week I write a column for the Yorkshire Post. This week’s column stopped me cold as I was writing it.

Most of you have spent years clicking on traffic lights, crosswalks, and fire hydrants to prove you are human. You did this hundreds of times. Possibly thousands. You found it mildly irritating. You assumed it served you.

It did not.

reCAPTCHA, launched by Carnegie Mellon professor Luis von Ahn in 2007, was ingenious. One word you solved proved your humanity. The other word, scanned from a book a machine could not read, was your unpaid labour. You transcribed it. You moved on. Google acquired the company in 2009 and pointed those millions of human answers at the New York Times archive, thirteen million articles back to 1851, then at Google Books. By 2011, the world’s literary heritage was digitised. By humans. Without a contract in sight.

Then the work changed. By 2012 the words were replaced by image grids taken from Google Street View. Every click told a machine what a traffic light looks like. Where a pedestrian crossing begins. Where a road ends. You were not passing a test. You were annotating the visual world for free.

You thought you were logging in. You were clocking in.

A 2023 study by the University of California, Irvine calculated that reCAPTCHA had extracted 819 million hours of unpaid human labour. Worth at least $6.1 billion in wages never paid. Among other things, it helped build Waymo, Google’s self-driving car business, valued at $126 billion as of February 2026.

Today reCAPTCHA v3 does not even show you the grid. It watches how you move your mouse. How you scroll. How long you pause between keystrokes. Whether you move in the jerky, slightly irregular pattern of a human or the smooth, precise arc of a script. Your behaviour is the training data. You cannot see the task. You cannot refuse it. You are not even aware it is happening.

It is not a test you pass. It is a tap left running.

This is not a test. It is a production system. And you are inside it.

There is a word for a system that extracts productive labour from people without the meaningful ability to refuse it. We have historically used that word carefully. 819 million hours makes it less abstract.

I have spent thirty years in this industry. I worked directly alongside the people who built this stuff in Silicon Valley. I advised governments on where it was heading. I did not know this was happening until recently.

I told my son. He shrugged.

“Yeah,” he said. “I just assumed that’s what they were doing.”

That exchange should stop us cold. I was oblivious. He took it for granted. No outcry. No class action. No parliamentary debate. Billions of hours of human labour extracted without a signature, a disclosure, or a penny in return.

We debate endlessly about what AI will do to us. But here is the truth nobody wants to say out loud: before the AI era has properly begun, we were already conscripted into building it. Without consent. Without payment. Without even noticing.

That was the warm-up act. The main event is worse.

Part Two: The Great Compression. You Trained Your Replacement.

What happened quietly on the internet is now happening loudly in the labour market.

reCAPTCHA was passive. You did not know you were doing it.

What has happened to the global software development industry over the last three years is active. Deliberate. And the people doing it were praised every step of the way.

The architect who automated himself out of existence

A senior backend architect. Two hundred and eighty thousand dollars a year. Six months spent building AI-enhanced developer tools for his team. He trained the junior developers on Cursor. Set up the Claude integrations. Automated the code reviews. Built, by his own account, something exceptional.

What he did not realise was that management was watching very carefully. Not his output. The AI’s output.

The platform he built could handle 73 per cent of the platform’s complexity with zero human oversight. The remaining 27 per cent was routed to a contractor in Romania at twenty-eight thousand dollars a year. The math was not complicated.

His manager pulled him aside. “Your AI tools work so well we don’t need the team anymore. Just the tools.” They kept two people out of eleven. He was not one of them.

His performance review, filed before the restructure, said: “Exceptional work on AI integration. You’ve revolutionised how we approach development.”

He automated himself out of existence and got praised for it.

The same story unfolded at Stack Overflow. The co-founder, on a package north of four hundred thousand dollars, was managed out after building the AI training pipeline that ingests developer conversations. His replacement was the pipeline he built plus an eighty-nine-dollar-a-month API subscription. The irony is suffocating.

These are not anomalies. They are the template.

The Great Compression

For two decades, the offshore model was the unquestioned infrastructure of global technology. The math was straightforward: why pay a US junior developer ninety thousand dollars when you could hire four developers in Bangalore for the same price? The result was the Pyramid. One expensive onshore architect managing dozens of offshore workers handling boilerplate code, manual testing, documentation.

Then came the Copilot trap.

Between 2023 and 2024, tools like GitHub Copilot and ChatGPT were rolled out to offshore teams under the banner of productivity improvement. It worked. Offshore developers used AI to finish tickets faster. Every accepted suggestion, every corrected bug, every documentation comment was captured as training data.

The offshore teams were teaching the models how to do their own jobs.

By 2025, the routine work that justified the offshore model, the 70 per cent that was repetitive and well-defined and easily specified in a Jira ticket, was exactly what the AI had become perfect at. Companies began realising that the pyramid model no longer required most of the pyramid.

The Pyramid collapsed into a Pillar.

Headcount was the product. Now the product is the system.

Today the model is called agent-shoring. Instead of sending work to people in other countries, companies send it to autonomous AI agents. A digital teammate costs twenty to one hundred dollars a month. Even the cheapest offshore developer costs two to three thousand dollars a month when management overhead and infrastructure are included.

The 70 per cent that was offshore is now automated. The 30 per cent that remains needs one senior person to oversee the agents. In many organisations, that person is not even in-country.

The economic stakes

This is not a technology trend. It is a labour market event. And it has already started.

India’s IT sector contributes close to 10 per cent of GDP. Software exports account for over two hundred billion dollars annually. More than five million people employed directly. The tech campuses, the apartment blocks, the schools, the entire consumer economy of Bangalore, Hyderabad, Pune: all of it premised on continued demand for human developers. Entry-level hiring has dropped approximately 30 per cent in 2026. The ladder has been removed from the bottom. You can still reach the top. You just cannot get on.

The Philippines built its middle class on call centres. Klarna’s AI assistant now handles the equivalent of 700 full-time customer support agents. That is not a rounding error. That is the signal.

Eastern Europe supplied a generation of nearshore engineering talent. EPAM, Globant, Endava: business models built on the premise that skilled developers in Poland, Romania, and Ukraine could undercut Silicon Valley. That arbitrage is closing. Not because those developers have become less skilled. Because the AI has become skilled enough.

The human cost of living is not moving to match the agent. And unlike previous waves of displacement, there is no obvious alternative runway. The new jobs are not being created at the rate or in the locations where the old ones are disappearing. The timeline is compressed. The geography is wrong.

What the stock market already knows

Investors understood this before most analysts did.

EPAM Systems, long regarded as the gold standard for high-end software engineering, saw a 16 per cent single-day decline in February 2026 on cautious organic growth guidance. The trajectory of broader IT services firms tells the same story: the S&P BSE IT index fell sharply in February as investors began questioning the moat of headcount-based revenue models entirely.

Globant, EPAM, Accenture, TCS and Infosys all felt the pressure. Some have pivoted effectively. Accenture, large and diversified enough to reframe itself as AI infrastructure, has held up better than most. The pure-play offshoring firms that still monetise primarily through the number of bodies in seats are being punished by the market with no sign of reprieve.

Valuation multiples for the sector have normalised to around 9.8 to 10.5 times EV/EBITDA, down from pandemic-era highs of 15 times or more. The “seat model” is being priced out of existence.

The market is not predicting disruption. It is pricing the disruption that has already arrived.

Part Three: Jensen Huang Drew the Line. Which Side Are You On?

Jensen Huang said something this week that should cut through every career conversation happening in British business right now.

The Nvidia CEO, speaking on the Lex Fridman podcast, did not hedge.

“If your job is the task, then you’re very highly going to be disrupted.”

Not might be. Not eventually. Very highly going to be.

If your job is the task, you are already late.

That single distinction between a job and a task is the most important career diagnosis anyone will hear this decade. If you show up every day to execute a repeatable process, you are the process. And the machine runs processes better than you. Faster. Cheaper. Without breaks. Without errors. Without a salary negotiation.

The moment your role can be written as a checklist, the checklist gets automated. Your desk gets cleared.

Huang did not stop at the diagnosis. He handed you the prescription in the same breath.

“If your job’s purpose includes certain tasks, then it is vital that you go learn how to use AI to automate those tasks.”

Your job includes tasks. But your job is not the tasks.

Your job is the judgment around them. The decisions. The context. The instinct for why the work matters and what to do when everything breaks. That stays human. Everything else gets handed to the machine.

And the person who hands it over first does not lose their job. They become more valuable than everyone still doing it by hand.

What Huang is really saying

Disruption warnings have been everywhere for a decade. People discount them. What makes this different is the specificity. Huang runs the company that makes the hardware inside every significant AI system on the planet. He is not guessing. He is reading his own order book.

And note what he did not say. He did not say the people who learn to use AI will be safe. He said the people who use AI to automate their tasks will be more valuable. That is a narrower, more demanding claim. Using AI passively, running prompts, generating first drafts, does not move you up the value chain. It makes you marginally more productive at the tasks being automated anyway. The move he is prescribing is active. You identify the repeatable components of your work. You hand them to the machine. You use the time to go deeper on the work that cannot be written as a checklist.

The accountant who automates data entry becomes the strategist. The marketer who automates reporting becomes the creative. The person who refuses to automate anything becomes the most expensive way to do the cheapest work. That line is drawing itself right now, in every organisation in the country.

What the shift actually looks like for software developers

The market for people who simply write code is contracting. The value has moved to those who design the workflow, audit the output, and own the judgement.

The new stack of skills that separates the replaceable from the indispensable:

Move from executor to orchestrator. Learn agentic IDEs, Cursor, Windsurf, Claude Code. The skill is no longer syntax. It is context engineering: knowing exactly what documentation, architectural patterns, and business rules to feed the AI to get a working result.

Move up to systems thinking. AI can build components. It cannot reason about why you should choose one architecture over another for a specific scaling constraint. That requires trade-off analysis, distributed systems knowledge, and organisational context. None of it can be specified in a Jira ticket.

Build vertical domain expertise. AI is a generalist. If you know everything about fintech compliance, or healthcare data edge cases, or legacy COBOL systems, you become the guardrail for the agent. That is not replaceable. That is the last mile.

Master the AI-native stack itself. RAG and vector databases. AI evaluation frameworks. Multi-agent orchestration tools like LangGraph and CrewAI. If you can build the system, configure its guardrails, detect when it is hallucinating, and fix its reasoning path, you are indispensable. If you can only use the system, you are its product.

Develop non-automatable human skills. Complex negotiation. Stakeholder management. The ability to sit in a room with a sceptical CEO and explain what is actually happening. AI cannot navigate office politics, cannot empathise with a frustrated client, cannot convince a board to change strategy. Engineers who understand UX and revenue are substantially more valuable than those who only understand the repository.

The six-month roadmap

Month one: stop coding like a human. Start steering like an operator. Use Cursor, Windsurf, Claude Code to manage intent across complex codebases. Take a legacy open-source project and refactor it using only agentic tools. Do not write a single line by hand.

Month two: learn how AI memory works. Semantic search, vector databases, RAG. Build something that connects a model to private data without hallucinations. The goal is understanding, not just usage.

Month three: build a squad. One agent writes code, one reviews it, one writes the tests. Create an autonomous bug-hunter that monitors an issue tracker and drafts patches. This is how production AI teams operate in 2026.

Month four: learn to prove a system is reliable. Hallucination detection. Evaluation-driven development. A CI/CD pipeline that runs accuracy tests on agent output. This is work almost no one in most organisations can currently do. That gap is your opportunity.

Month five: understand the probabilistic stack. Token cost, latency, agentic observability. These are systems that do not behave the same way twice. Learning to run them in production is where most people stop. Do not stop.

Month six: build something real. A vertical solution where human judgement is the final gate. A tool that solves a high-value problem that AI cannot solve without you. This is the job description of the person who survives.

Huang is right that dislocation is coming. He is also right that the window to move is still open.

It is not wide. It will not reopen.

📊 The Sunday Signal Tech and AI Layoff Tracker: Week 14

Sourced from Layoffs.fyi and LayoffHedge.com. All figures current to 28 March 2026.

Total 2026 YTD: approximately 74,500 Added this week: approximately 8,800 LayoffHedge sentiment: Extremely Bullish on AI-native restructuring

This is what the transition looks like in numbers.

The narrative has shifted this week from efficiency to liquidation. Companies are cutting human capital to fund the compute clusters required to stay competitive. Eight thousand eight hundred roles in seven days is not a correction. It is a structural event.

The signal this week: the margin-labour swap

For the first time, companies citing AI-native replacement as their primary layoff reason appear to be receiving a median stock price premium of around 9 per cent compared to peers citing macroeconomic headwinds, based on tracker-based estimates. The market is no longer looking for cost-cutting. It is looking for human-to-agent conversion. Employees are being treated as technical debt. Agents are being treated as productive assets.

The KPMG threshold

The KPMG layoffs represent a critical moment for professional services. Eight hundred roles globally in audit and advisory is not large by absolute measure. It is significant for what it signals.

The Big Four model relied on high turnover. Mid-level roles would be vacated as people moved to industry, creating space for the cohort below to move up. That churn has stopped. Because mid-level roles are disappearing from the market due to AI saturation, people are not leaving. The talent is trapped in place. KPMG is cutting because it can no longer wait for natural attrition. The audit agent has arrived and the people it replaces have nowhere else to go.

Watch Deloitte and PwC for announcements in the coming weeks. The KPMG decision will echo.

Oracle and the infrastructure trade

Oracle confirmed approximately 5,000 roles across global enterprise divisions, per tracker-based estimates. The analysis from LayoffHedge frames this as a direct liquidity play to fund its three-hundred-billion-dollar cloud partnership with OpenAI. The stock rose because investors read it correctly: Oracle is selling five thousand salaries to buy GPU clusters.

The logic is brutal and it is now mainstream. Labour is being repriced as operating expenditure. Compute is being repriced as capital investment. Every major institution is running the same calculation.

Next week: the risk heatmap

The Capex Gap analysis for week 14 identifies five companies as highest probability for layoff announcements in the coming seven days. Microsoft is critical. The Stargate funding timeline is creating liquidity pressure in non-AI Azure and hardware units. Salesforce is critical. Agentforce 2.0 is eliminating the sales rep roles that the company simultaneously markets the product to. Deloitte, Alphabet, and ServiceNow complete the watchlist.

In 2026, labour is the new interest expense. The more humans a company carries, the more expensive its balance sheet looks to an AI-first analyst.

🚀 Final Thought

reCAPTCHA extracted six billion dollars of human labour without paying a penny. The offshore model extracted three decades of expert knowledge and fed it into the training data that made that expertise obsolete. The developers who built the best AI tools got managed out because the tools worked.

The machine did not take your job.

You gave it the training data, the workflow, and the permission.

The people who understand the distinction between a job and the tasks within it, between execution and judgment, between being the product and designing the system, those people are not being displaced.

They are being promoted.

Most people will realise this too late.

Until next Sunday, David

David Richards MBE is a technology entrepreneur, co-founder of Yorkshire AI Labs, and a weekly columnist for the Yorkshire Post. The Sunday Signal is published every Sunday at newsletter.djr.ai

Listen to this issue as a podcast. Available on Spotify, Apple Podcasts, and YouTube.

65,700.

That is how many roles have been eliminated or marked for elimination in the first eleven weeks of 2026. Not over a year. Not over a cycle. Eleven weeks.

AI is not starting at the bottom of the labour market. It is starting in the middle. And the pace is not slowing. It is accelerating.

This week has three stories, and they belong together.

The first is the Yorkshire Post column published this week. The Full Monty was about Sheffield steelworkers who lost the thing that defined them. This time it is bankers, analysts, lawyers and accountants facing the same structural shock. The white-collar professional class assumed its credentials would protect it. They will not.

The second story is the one that changes the terms of the debate entirely. Jeff Bezos is in early-stage talks to raise up to $100 billion for a fund designed to buy established manufacturers and overhaul them with AI. Chipmaking. Aerospace. Automotive. Defence. The people who told you that physical, hands-on work was safe from this wave were reasoning from the last one. Bezos just confirmed, with the most ambitious private investment thesis in history, that they were wrong.

The third story is what happens when the office and the factory go at the same time. The economics, the social fracture, and the governance crisis no government is prepared for.

The Sunday Signal Tech and AI Layoff Tracker returns below, with a significant new entry from the banking sector that makes the direction of travel impossible to ignore.

You can also listen to The Sunday Signal as a weekly podcast using a voice cloned from my own voice, a demonstration of the same AI capabilities I write about. Listen on Spotify, Apple Podcasts, and YouTube.

Subscribe now

Story One: The Full Monty Was About Steelworkers. This Time It’s Bankers.

This week’s Yorkshire Post column, adapted for The Sunday Signal.

The remake opens in 2036.

A glossy corporate film flickers on screen. The City of London at sunrise. Glass towers glowing. Young analysts moving at speed between meetings and markets. Screens flashing. Deals closing. A voiceover promises that for those with talent and drive, finance will always reward human judgment.

Then it stops.

Silence.

The trading floor is empty. Screens are dark. The work has not moved to New York or Singapore. It has not been outsourced. It has simply dissolved into software.

In the corner, two former bankers in scuffed Loakes are trying to wrench a server from its rack, hoping to strip copper from the very machines that replaced them.

It sounds far-fetched. Until you remember we have seen it before.

The original Full Monty was not about abstract economic forces. It was about Sheffield steelworkers who lost the thing that gave them identity, pride and purpose. When the work went, it did not just take wages. It took status, structure and meaning.

For a generation, that kind of shock was understood as a working-class problem. Something that happened to people who worked with their hands. White-collar professionals assumed they were insulated. Degrees, offices and credentials were meant to protect them.

That assumption is now breaking down.

In the first eleven weeks of 2026, major technology companies have announced or are developing cuts totalling approximately 45,724 roles, based on tracker-aggregated estimates from Layoffs.fyi and RationalFX. That number is already out of date. By the time you read this, it has crossed 65,700. The developing stories, including HSBC and the heavily reported Meta cuts, suggest the real figure is still climbing.

Behind the headlines, something more important is happening. Roles are not being replaced. Teams are not being rebuilt. The first rung of the career ladder is quietly being removed.

AI is not starting at the bottom of the labour market. It is starting in the middle. Software, finance, business operations and law dominate usage. And the gains increase with complexity. The more educated and valuable the task, the more it can be accelerated.

The real shift is not speed. It is what happens to the job itself.

When AI takes over drafting, analysis, coding or planning, what remains for the human is the lower-value residue. Coordination. Oversight. Exception handling. The job does not vanish overnight. Its intellectual core erodes.

The steel industry followed exactly this path. Technology did not simply replace workers. It redefined the work, broke skilled roles into simpler tasks, and reduced the economic value of labour over time.

That same dynamic is now moving through the office.

The clearest signal is not layoffs. It is hiring. Firms are no longer creating the entry-level roles that once trained the next generation. The ladder is not being climbed. It is being removed.

And when you remove the first rung, you do not just lose jobs. You break the system that produces experience, judgment and expertise.

When deindustrialisation swept through cities like Sheffield, the people most affected did not have the influence to force a rapid response. The adjustment came late, and often in the form of retraining programmes that never quite matched the scale of the problem.

The professional class will insist its work is different. It always does.

Right up to the point where it is not.

Story Two: Bezos Has $100 Billion and He Is Coming for the Factory

The assumption that physical work is safe just collapsed.

There is a comfortable argument that many in manufacturing still reach for.

Software jobs, yes. Finance, perhaps. But you cannot automate a factory. You cannot AI your way through a chipmaking line. The physical world is different.

It is not different.

Jeff Bezos is in early-stage talks to raise up to $100 billion for a new investment vehicle. The pitch documents, as reported by the Wall Street Journal on 19 March and confirmed by Reuters, the New York Times, Forbes and Bloomberg, describe it as a “manufacturing transformation vehicle.” The strategy is brutally simple: buy established manufacturers in chipmaking, defence, aerospace and automotive, then overhaul them with AI. Not gradually. Decisively.

The fund operates alongside Project Prometheus, Bezos’s secretive AI startup co-founded with former Google X executive Vik Bajaj. Prometheus is not building a chatbot. It is building what its team calls Physical AI: models that understand and simulate real-world physics, engineering and manufacturing processes. Airflows. Stress points. Thermal dynamics. The things that govern how physical objects are made, tested and repaired.

Prometheus launched in November 2025, received $6.2 billion in initial funding, and Bezos serves as co-CEO. His first operational executive role since stepping down at Amazon in 2021. That detail tells you everything about how seriously he takes this.

The $100 billion fund would give Prometheus something a laboratory cannot provide: real factories, real supply chains, real workforces to transform.

The scale is historic. The only comparable private fund ever raised is SoftBank’s original Vision Fund in 2017, also $100 billion. But SoftBank bet on startups. Bezos is betting on established industrial businesses and replacing their operating model with AI. That is not a venture bet. That is a restructuring of the physical economy.

Bezos has been personally on the roadshow. Sovereign wealth funds in the Middle East. Asset managers in Singapore. No money has closed yet. But the ambition is not a rumour.

The direction is clear: fewer people, more machines, longer operating hours, and a shift from labour to systems. Not incremental improvement. Structural replacement.

Consider what this means for the jobs we assumed were protected by their physicality.

Manual assembly: replaced by multi-purpose robotic arms running continuous shifts. Quality control: replaced by AI computer vision that detects microscopic defects faster and more consistently than any human inspector. Logistics: managed by autonomous mobile systems. Maintenance: directed by AI copilots overlaid on the physical hardware through augmented reality, reducing the need for years of accumulated trade expertise.

The roles that survive are systems engineering and orchestration. High-demand. High-skill. Numerically small.

Last week I described the one-person unicorn: companies generating extraordinary revenue with workforces of dozens rather than hundreds. Cursor, Midjourney, Gamma. That model was about software. Bezos is extending it to chipmakers, aerospace manufacturers and automotive plants. The same logic. A different industry. A bigger disruption.

Previous waves of automation replaced specific machines. A lathe. A press. A welding arm. Everything else remained human. Project Prometheus is designed to replace entire factory ecosystems. Not the weld. The whole floor.

There are 425,000 unfilled manufacturing roles in the United States right now that require AI literacy. The jobs that will survive this wave already exist. The problem is that the current workforce does not yet have the skills to fill them, and the retraining infrastructure to close that gap does not yet exist at anything approaching the scale required.

The people who told you the physical world was safe were reasoning from the last wave. This is a different wave.

Story Three: The Robot Tax Is Coming. And It Will Not Be Enough.

The economic and social consequences of simultaneous displacement.

Governments are already talking about a robot tax. A levy on automated output. A charge on compute. A payroll equivalent paid by companies that replace workers with machines. The debate is live in Brussels, in Sacramento, in Westminster. It sounds like a solution.

It is not a solution. It is a symptom. A sign that the fiscal foundation of the modern state is cracking and policymakers are reaching for the nearest available tool.

Here is why it matters, and why it will not be enough on its own.

Every previous wave of automation has had a partial quality to it.

The loom replaced hand-weaving but left the mill. The assembly line replaced craftsmen but created line operators. The spreadsheet replaced clerks but created accountants. In every case, displacement was sector-specific. New roles absorbed the shock, eventually, if imperfectly.

What Bezos is building, alongside the AI-native software companies I described in recent issues of this newsletter, is something without clear historical precedent. Simultaneous displacement across both the white-collar professional economy and the physical manufacturing economy. Not sequentially. At the same time.

The economics are stark.

Manufacturing is the last refuge of labour arbitrage. For decades, if wages rose in one country, production moved to another. Vietnam. Indonesia. Mexico. The competitive advantage of low-cost human labour has kept hundreds of millions employed in the developing world and provided a brake on wage deflation in the West.

That brake is about to be released.

If an AI-run factory in Yorkshire costs the same to operate as one in Vietnam, because the primary input is no longer human labour but electricity and chips, production comes home. Not because of trade policy or political will. Because of arithmetic. And when that happens, the entire logic of globalisation begins to unwind.

This creates something economists are only beginning to model. A manufacturing renaissance in wealthy nations, accompanied by severe economic disruption in developing ones. Countries that have spent fifty years building export manufacturing capacity will find that advantage evaporate. That is not a domestic British story. It is a global destabilisation event.

At home, the social impact follows the fault line I have been tracking in this newsletter since Issue #40.

The High-Agency tier thrives. Engineers, AI orchestrators, systems architects. People who direct these systems rather than operate within them. Their leverage is extraordinary. A small team managing a Prometheus-driven production line can outperform a workforce of hundreds.

The Displaced Tier faces something harder than unemployment. It faces irrelevance. A factory foreman with thirty years of floor knowledge. A bank analyst with a first from Durham. The AI replacing them is not approximating. It is outperforming, consistently, at a fraction of the cost.

There is a governance crisis buried in these numbers that governments are not yet confronting at the scale the problem demands.

If forty to fifty per cent of manufacturing employment is automated alongside significant white-collar displacement, the income tax and national insurance base fractures. The modern state is built on taxing labour. Remove labour, and you remove the foundation. Everything else is detail. The debate about a robot tax, a compute levy, or a tax on AI-generated output is not a theoretical policy question. It is a fiscal emergency arriving in slow motion.

And underneath the economics is something the data cannot capture.

Sheffield in the 1980s was not just an unemployment story. It was a meaning story. Work is not merely income. It is identity, structure and purpose. The Full Monty was not about poverty. It was about what men become when the thing that defined them is taken away.

We are about to run that experiment again. Simultaneously. Across every class, every sector, every region.

The handloom weavers of 1810 could see the machines arriving. They understood what would happen. Their response has been misrepresented by history as mindless rage against progress. It was not. It was a demand that progress should not come at the cost of people who had no other options. That demand was ignored.

We should be very careful about ignoring it again.

The Sunday Signal Tech and AI Layoff Tracker

Updated weekly. Sources: Layoffs.fyi, TrueUp.io, RationalFX, Reuters, Bloomberg. As of Sunday 22 March 2026.

Total roles eliminated or reported for elimination in 2026: approximately 65,700.

Up from 45,724 as of 14 March. The most significant new entry this week is HSBC, which disclosed plans to reduce its workforce by up to 20,000 roles, explicitly citing AI automation in back and middle office functions. This is the first globally systemically important bank to name AI as the mechanism of reduction in formal planning documents. It will not be the last.

Meta and Oracle remain in developing status, with Meta in particular carrying heavy internal confirmation signals for cuts of 15,800 to 16,000 roles tied directly to the reallocation of its $135 billion AI capital expenditure programme.

The pattern is consistent across every confirmed entry. Every company that has made an explicit cut tied to AI has seen its share price rise or hold. Block surged 25 per cent on the day of its announcement. WiseTech gained 11 per cent. Atlassian rose between 2 and 4 per cent. The market is not mourning these jobs. It is pricing in the leverage.

HSBC is a different signal. Layoffs at a software company, however large, can be rationalised as sector-specific. A globally systemically important bank, operating across 62 countries, telling regulators and investors that AI will eliminate one in ten roles, explicitly, as a plan rather than a possibility, is a different category of event. Finance was meant to be complex enough to resist this. HSBC has now answered that assumption directly.

🚀 Final Thought

Last week I wrote about the one-person unicorn. The company that generates a hundred million in recurring revenue with a workforce measured in dozens. That story was about software. This week, Bezos has extended the same thesis to steel, chips, aerospace and automotive.

The physical world is not different. It is just next.

The Luddites did not stop the loom. But the political pressure they generated, the social disruption they created, eventually produced the Factory Acts. Child labour protections. Workplace safety law. The early architecture of a social contract between capital and labour. The machine won. But it forced a renegotiation.

What we need now is not a fight against AI. It is a serious, adult, politically difficult conversation about what the social contract looks like when the machine can do the job better than the human. About who owns the productivity gains. About what replaces the purpose that work once provided.

That conversation is not happening at the scale the problem demands.

It will happen eventually. The only question is whether we start it now, while choices still exist, or wait until the factory floor is dark and the trading floor is empty, and two former bankers are stripping copper from a server rack.

Until next Sunday, David

David Richards MBE is co-founder of Yorkshire AI Labs, a weekly columnist for the Yorkshire Post, and the author of The Sunday Signal. Published every Sunday at newsletter.djr.ai.

Listen to this week’s episode on Spotify, Apple Podcasts, and YouTube.

The Bottom Line Up Front

45,724.

That is the number of tech and AI-related roles eliminated or reported for elimination in 2026 so far, according to tracker-based estimates from RationalFX and Layoffs.fyi. We are not yet through the first quarter.

Twenty-five employees. One hundred million dollars in recurring revenue. Achieved in twelve months, with no marketing budget, by a team of engineers and founders that would not fill a typical company’s finance department.

That is what Cursor looked like when it hit a milestone that once implied an entire mid-sized company. That is not a curiosity. It is a warning.

This week’s issue has three parts. First, the Yorkshire Post column: the one-person unicorn is coming, and most business leaders still do not grasp the scale of the change. Second, the mechanics: how AI-native companies are replacing headcount with agents in engineering and go-to-market. Third, the stack: the tools making this possible, and why the cost of testing a serious company idea has collapsed.

And from this week, The Sunday Signal adds a permanent new section: The Sunday Signal Tech and AI Layoff Tracker.

The numbers demand it. Atlassian cut 1,600 roles on 11 March. Block cut over 4,000 in late February. WiseTech cut 2,000. Amazon is rolling through 16,000 corporate positions. In every confirmed case, management is making the same strategic bet: fewer people, more automation, more AI leverage. In every confirmed case, investors responded by pushing the share price up. Block surged 25 per cent on the day. WiseTech jumped 11 per cent. Atlassian rose between 2 and 4 per cent.

The market is not mourning these jobs. It is pricing in the leverage.

I have been building to this section since Issue #40 on 8 February. The data has been accumulating week by week. It now has its own space, every Sunday, for as long as this wave continues.

You can also listen to The Sunday Signal as a weekly podcast using a voice cloned from my own voice, a demonstration of the same AI capabilities I write about. Listen on Spotify, Apple Podcasts, and YouTube.

Subscribe now

Story One: The One-Person Unicorn Is Coming

This is the column I wrote for the Yorkshire Post this week.

Twenty-five.

That is roughly how many employees Cursor, the AI-powered coding company, had when it crossed one hundred million dollars in recurring revenue. Achieved in twelve months, with no marketing spend, through a product so good that developers simply bought it themselves.

I have spent three decades in technology. I am not easily impressed by revenue milestones. That number stopped me cold.

To understand why, remember what £100 million used to imply. A conventional software company reaching that threshold would typically employ between 400 and 800 people. Sales teams chasing pipeline. Customer success managers defending renewals. Marketing operations tracking attribution. Layers of middle management coordinating it all. I have run businesses at that scale. I know what that machinery costs to assemble and to maintain.

Cursor made much of it unnecessary.

And it is not alone.

ElevenLabs closed last year with more than $330 million in ARR. Sierra, founded by Bret Taylor, hit $100 million ARR in seven quarters. Midjourney has shown that an AI-native company can reach extraordinary revenue with a team most corporates would still regard as dangerously lean. The exact ratios vary. The structural shift does not.

That is not incremental improvement. It is structural change.

These businesses were not built as conventional organisations. They were built as automated loops.

Work that once required 200 engineers is now directed by a handful of senior people orchestrating AI agents that write, test and refactor code. Customer support that might have required 100 staff is handled by reasoning systems that check usage logs, interrogate billing records and issue refunds without human intervention. Midjourney does not operate a traditional sales force. It does not need one. Credit card in, product out.

Revenue scales. Headcount barely moves.

The underlying shift is agentic AI. Earlier AI systems summarised and suggested. Useful, but passive. The current generation executes. It builds the feature, runs the tests, deploys the update and handles the complaint. It chains decisions together and improves through feedback. When you map that capability onto a company’s cost base, it does not look like a new tool. It looks like a redefinition of what a company actually is.

For thirty years, growth meant hiring. Hiring meant cost. Cost meant capital. That logic shaped everything from venture funding to office design. Now we are watching leverage detach from labour.

My view is that we are perhaps eighteen months from the first genuine one-person unicorn. A founder directing a fleet of AI systems to build, market and operate a billion-dollar business.

That sounds implausible. So did the idea that a global software company could reach nine figures in recurring revenue with fewer than twenty-five people.

When I started in this industry, launching a product required racks of servers, specialist contractors and months of burn before the first customer invoice. Cloud computing collapsed infrastructure. Modern frameworks reduced development friction. No-code tools shrank teams further. Each wave removed a reason to hire.

Agentic AI may remove the assumption that you need a team in the traditional sense at all.

That should concentrate minds.

Every business leader faces the same uncomfortable question. Are you building for the world that existed five years ago, or the one arriving at speed? The unicorns of the next decade will not be defined by the size of their workforce. They will be defined by the scale of their leverage.

The army, in many cases, has already become the liability.

Story Two: This Is How They Are Actually Doing It

The question everyone asks next is straightforward. How?

The answer matters because these companies are not using AI as a feature bolted on to a conventional business. They have built what the industry now calls an AI-native architecture. The distinction is fundamental. In a traditional company, the software is a tool and the people do the work. In an AI-native company, the agents do the work and the people set the direction.

The result is that human employees have become the bottleneck, not the engine. A small team of senior architects. A fleet of agents executing. No meetings required for the agents to coordinate.

Here is how it plays out in the two departments that consume the most headcount in any growing technology company.

Software Development: The Agentic Fleet

In a traditional company, developers write code. In an AI-native company, developers direct agents that write code.

The directional numbers are striking. Reports from inside companies like Cursor suggest that a significant proportion of the pull requests merged into their product are now created by autonomous AI agents running in the cloud. A small team of architects defines the problem, sets the constraints and spins up fleets of agents working in parallel. Far fewer junior engineers are needed for repetitive, structural work.

When a bug is detected, an AI agent automatically creates a virtual machine, reproduces the bug, writes the fix, and presents a recording of the fix to the human architect for a thirty-second approval.

The industry has coined a term for the developer working this way: vibe coding. A developer describes a high-level intent: “make the checkout flow handle crypto and send a Slack alert”. The AI then manages the thousands of lines of code required to make it happen. The developer never touches the implementation.

The revenue trajectory tells you what this produces. A traditional SaaS company at £100 million ARR employs around 500 people, generating roughly £200,000 per head. Cursor has crossed $1 billion in annualised revenue at a scale of headcount that would once have looked absurdly small. Lovable reached $100 million ARR within eight months of launch and has continued climbing fast.

The point is not the exact spreadsheet ratio. The point is that AI-native companies are scaling output far faster than headcount. The difference is not better engineers. It is a different organisational architecture.

Human replacement estimate: The agentic fleet replaces roughly 150 to 180 junior and mid-level engineers who would otherwise be responsible for boilerplate code, manual testing and bug fixing.

Marketing: Programmatic Growth

Traditional marketing is people-heavy because it requires constant content creation, manual outreach and complex attribution tracking. Lean AI companies have replaced this with what researchers call autonomous growth loops.

Three techniques dominate.

AEO: Answer Engine Optimisation. One or two people use AI to ensure their product is the top recommendation when a potential customer asks ChatGPT, Gemini or Perplexity for advice. They do not write blog posts for humans. They build knowledge graphs for AI. This is rapidly becoming the 2026 version of SEO, and most marketing teams have not yet understood it exists.

Content atomisation. Instead of a social media team, agents take one piece of source content (a founder’s technical talk, say) and automatically produce 50 or more localised videos, posts and ads, each adapted for different markets and platforms. A single piece of original thinking becomes a global content operation without a single additional hire.

The invisible sales force. Midjourney and Cursor have scaled to hundreds of millions in revenue with no traditional sales force. For companies that do need outbound, AI agents now monitor the web for trigger events (a competitor raising prices, a target company hiring aggressively) and automatically generate personalised demos or sequences for every affected prospect. No SDR required.

Human replacement estimate: This replaces 40 to 60 staff across marketing operations, SEO, content creation and sales development. One GTM engineer now does the work of an entire traditional marketing department.

The combined picture is stark.

Story Three: The Exact Tools They Use

The research is now deep enough to be specific. Based on verified implementations documented across a16z Speedrun, SaaStr, Growth Unhinged and company disclosures, here is what ultra-lean AI companies actually deploy, and what it costs.

The single most important principle before the tools: the most successful lean companies consolidate aggressively. Five to ten core tools rather than thirty. Every additional tool creates coordination cost and cognitive load. The savings are reinvested in AI agent quality.

The Core Stack

Sales and prospecting

The dominant documented pattern is a three-tool combination: Apollo for prospecting (270 million verified contacts, free starting tier) → Clay for enrichment → Instantly for delivery.

Clay itself has crossed $100 million ARR and sits at the centre of nearly every documented lean GTM stack. It is to modern GTM what Salesforce was to the previous era. Its waterfall enrichment sequentially queries 100 or more data providers to maximise contact coverage. OpenAI’s own GTM team used Clay to double contact coverage from 40 per cent to 80 per cent. Rippling used it with hiring data triggers and doubled cold email performance year-on-year.

One GTM engineer running this combination can manage thousands of personalised prospects per week.

CRM

Attio (free to start, £29 per user per month) is the documented recommendation from the a16z Speedrun minimum viable stack. HubSpot free tier for maximum accessibility and zero commitment.

Content and marketing

ChatGPT and Claude serve as foundational writing tools. Midjourney, Runway and Canva AI combined eliminate the need for creative hires entirely. Typefully handles lean social publishing on X and LinkedIn.

Customer support

Intercom Fin has become the default AI support agent for AI-native companies. Fin (powered by Claude) resolves 51 per cent of queries on average without any human involvement. Gamma uses it to resolve 75 per cent of queries end-to-end for 50 million users. Anthropic itself chose Fin over building its own support agent.

Discord runs alongside Fin as the community hub. Midjourney built its entire business on Discord: 2.7 million members serving simultaneously as distribution channel, support platform and product interface.

Product analytics

PostHog has emerged as the consolidation play, replacing Amplitude, LaunchDarkly, Hotjar and Segment in a single platform. Free tier includes one million events per month. Documented users include Supabase, Lovable and ElevenLabs.

Automation backbone

n8n (178,000 plus GitHub stars, self-hostable, from £50 per month) is the standout tool for lean teams. It handles the workflow automation that once required a dedicated DevOps team. One Montreal startup documented using n8n to build an automation backbone typically requiring three or four dedicated operations staff.

AI agents and orchestration

For developer-led orchestration, LangGraph handles complex agent workflows with cycles, conditionals and state persistence. CrewAI ($18 million Series A, 100,000 plus agent executions per day) uses a role-based model for multi-agent coordination. For no-code orchestration, Lindy.ai positions as “AI employees” with pre-built templates for sales, support and recruiting, starting at £29 per month.

What the Whole Thing Costs

For under roughly £3,000 a month, a founder can now assemble a surprisingly capable commercial and operating stack. That does not guarantee a £100 million business. It does mean the cost of trying has collapsed.

The final insight from the research comes from Lenny Rachitsky: your job as a founder is to become world-class at one primary growth engine before layering anything else on top. The tool list above is not a shopping list. Pick the single motion that fits your business. Build it. Own it. Only then add the next layer.

📊 New Weekly Section: The Sunday Signal Tech and AI Layoff Tracker

Since Issue #40 on 8 February, I have been tracking the acceleration of tech layoffs. Issues #41, #42, #43 and last week’s #44 each added more data. The scale now demands a permanent weekly section. From this week, the Tracker runs every Sunday.

In the first eleven weeks of 2026, major tech companies have announced or are developing cuts totalling approximately 45,724 roles, according to tracker-based estimates from RationalFX and Layoffs.fyi. The unverified and developing stories suggest the real figure may be considerably higher.

But the raw totals are the least interesting part of the story. The patterns are what matter.

Pattern one: the AI pivot premium is real

The market is explicitly rewarding companies that announce they are cutting people to fund AI investment.

Atlassian cut 1,600 roles (10 per cent of staff) on 11 March and the stock rose approximately 4 per cent. Investors cheered the pivot from growth-at-all-costs to AI efficiency. WiseTech Global cut roughly 2,000 roles (30 per cent of staff) to replace manual coding with AI-native workflows. Share price up 11 per cent.

Block went furthest. Roughly 4,000 roles eliminated in a massive shift to AI-augmented lean teams and to reverse what the company described as the 2018-to-2024 bloat. The result was one of the largest single-day market cap gains in Block’s history, up approximately 25 per cent.

The signal is clear. Investors are not mourning the jobs. They are pricing in the leverage.

Pattern two: the SaaS crisis is emerging

Not every company is being rewarded for its AI pivot.

Workday cut 400 roles in customer operations in February. The stock is down around 40 per cent year to date. The reason is not the number of cuts. The reason is investor fear that agentic AI will not just assist Workday’s enterprise software. It may make the software itself unnecessary.

If developers stop writing structured HR workflows and instead describe what they want to an agent (the vibe coding model), then the value of a heavyweight enterprise HR platform collapses. Investors are beginning to ask whether tools like Workday remain as indispensable in an agentic world as they looked in the last software cycle.

This is the deepest structural risk I have been writing about since Issue #40. The AI disruption is not symmetrical. Companies building AI infrastructure are gaining. Companies that were built on the assumption of human cognitive labour are the ones in danger.

Pattern three: rumour versus reality

Not every number circulating in tech forums is confirmed.

Meta has reports of potential cuts affecting more than 20 per cent of staff circulating internally, with no formal announcement and no SEC filing to date. Oracle reports suggest potential large-scale cuts are developing; no official confirmation has been given. Treat both as directional signals until they appear in a regulatory filing, not as settled facts.

Amazon is confirmed. 16,000 corporate role reductions, framed explicitly as management flattening and a reallocation of $37 billion to AI compute and infrastructure. That one is in the numbers.

Pinterest cut 675 roles (approximately 15 per cent of staff) in a January/February cycle to pivot from a high-cost social platform to an AI-driven discovery engine. Stock down 3.5 per cent on the initial reaction, with recovery building as AI-targeting optimism grows. Autodesk cut approximately 1,000 roles (7 per cent) to reorganise around AI-driven generative design tools. eBay has confirmed approximately 800 roles, 9 per cent of full-time staff, in a strategic realignment.

Figures and market impact data are based on reports available as of 14 March 2026. Tracker estimates sourced from RationalFX and Layoffs.fyi. Developing stories without WARN notices or SEC filings remain unverified. Stock changes are immediate short-term reactions.

🚀 Final Thought

The army, in many cases, has already become the liability.

The one-person unicorn is not a thought experiment. It is an engineering problem. And several people are very close to solving it.

When I started in this industry, launching a product required racks of servers, specialist contractors and months of burn before the first customer invoice. Cloud computing collapsed infrastructure. Modern frameworks reduced development friction. No-code tools shrank teams further. Each wave removed a reason to hire.

This wave may remove the assumption entirely.

The question is not whether this is coming. It is whether you are on the right side of the shift.

David Richards MBE is the co-founder of Yorkshire AI Labs. He writes a weekly column for the Yorkshire Post and is a regular commentator on technology, business and the future of work.

The Sunday Signal is produced as a weekly podcast using a voice cloned from David’s own voice, a demonstration of the same AI capabilities he writes about. Listen on Spotify, Apple Podcasts, and YouTube.

Until next Sunday, David

The Bottom Line Up Front

Two weeks ago I wrote about the historical pattern. The handloom weavers of Yorkshire. The structural logic of technological displacement. The way skilled professions convince themselves they are immune until the moment they are not.

Last week the financial markets began to confirm that pattern. IBM fell sharply after warning that AI will reshape parts of its consulting workforce. Atlassian has lost more than 70 per cent of its value over the past year. Block cut roughly 40 per cent of its workforce and investors rewarded the decision. These are different companies in different sectors, but they point in the same direction. The cost of producing cognitive output is falling, and markets are beginning to price what that means for every business built on human expertise as its primary input.

This week the academic evidence arrived to match the market signals.

On 5 March Anthropic published the most rigorous labour market analysis yet of what AI is actually doing to employment. Not what it could theoretically do in some distant scenario, but what it is doing in real professional settings today, tracked across millions of workplace interactions.

The headline finding appears reassuring. There has been no sudden collapse in employment. The labour market has not broken.

Do not be comforted by that.

The real signal sits deeper in the data. It is the sort of signal that appears quietly in graphs years before it erupts into a political crisis. The sort of signal the handloom weavers saw in the early nineteenth century and understood perfectly well, even if they lacked the power to stop it.

This issue looks at three positions. The bull case for abundance. The bear case for decoupling. And my own view on which outcome is more likely. It is not a fully optimistic view and it is not a catastrophist one. It is structural. And it should make you uncomfortable wherever you sit in this debate.

If you prefer to listen rather than read, each issue of The Sunday Signal is also available as a podcast. Perfect for the commute, a walk, or time on the road. You can find it on Spotify Youtube or Apple Podcasts.

Subscribe now

Part One: The Gap That Is About to Close

What the Anthropic Study Actually Found

Anthropic introduced an important concept into the debate about AI and employment. They call it observed exposure.

Previous studies tended to focus on capability. Researchers examined what large language models could theoretically do, compared those capabilities with occupational task databases, and then estimated the percentage of jobs that might be at risk.

The problem with that approach is that it confuses potential with deployment. A tool may be capable of performing many tasks, but that does not mean it is actually being used to perform them. The capability assessment tells you the theoretical ceiling. It does not tell you where you are on the way to it.

Anthropic approached the question differently. Instead of asking what AI could do, they asked what it is doing. By analysing millions of real conversations with Claude in professional environments they mapped which occupational tasks were genuinely being performed by AI right now, at scale, in actual workplaces.

The results are striking.

Computer programmers already have 74.5 per cent of their tasks covered by AI systems. Customer service representatives sit at 70.1 per cent. Data entry clerks at 67.1 per cent. Medical records specialists at 66.7 per cent. Financial analysts at 57.2 per cent. Paralegals and legal assistants at 53.8 per cent.

Pause for a moment on the financial analyst figure. More than half of the working tasks performed by financial analysts are already being handled, at least in part, by AI systems in real workplaces today. Not in a laboratory. Not in a proof of concept. In the field.

But the most important number in the study is not the level of current deployment. It is the gap between deployment and capability.

Across every occupational category analysed, the theoretical capability of AI remains significantly larger than its current real-world usage. The spider chart in the report makes this visually obvious. A small red shape representing actual usage sits inside a much larger blue shape representing what the systems could already do.

The gap is large. And it is closing.

This is the signal that the reassuring headline about stable employment numbers obscures. We are not looking at the aftermath of an AI disruption. We are watching the early phase of one. The adoption curve has not peaked. It has not even reached its steepest point yet. The observed exposure figures being discussed this week may look like the floor in two years, not the ceiling.

AI does not need to reach its theoretical maximum capability to produce structural disruption. It simply needs to close the gap faster than the labour market can adapt.

On current trajectory that is exactly what appears to be happening.

The Youth Hiring Signal

Buried in the Anthropic data is an early warning signal that has received far less attention than it deserves. It should be the most discussed finding in the report.

Hiring of workers aged twenty-two to twenty-five in AI-exposed occupations has fallen by 14 per cent since the launch of ChatGPT in November 2022. Not unemployment. Hiring.

Companies are not firing experienced professionals en masse. Instead, they are quietly choosing not to create the entry-level roles that once absorbed new graduates.

That distinction matters enormously, and it is worth understanding precisely why.

Professional careers have traditionally been built on an apprenticeship model. Young analysts built spreadsheets and prepared briefings. Junior lawyers reviewed documents and drafted routine correspondence. Graduate developers maintained legacy code and fixed bugs. The work was repetitive and often tedious, but it served a function that went well beyond the immediate output.

It allowed each generation to accumulate the practical experience that eventually produced senior professionals.

The entry-level role was never just a job. It was the mechanism through which expertise was transmitted from one generation to the next. You could not become a senior partner without having been a junior associate. You could not lead due diligence without having spent years doing the groundwork of due diligence. The repetitive work at the bottom was not a waste of talent. It was the foundation on which judgment was built.

Remove the entry-level role, and you do not simply eliminate a job. You remove the first rung of the professional ladder. A generation of graduates cannot become the experienced partners, analysts and architects of 2035 if the early career positions that generate that experience no longer exist.

This kind of structural damage does not appear immediately in unemployment statistics. It does not generate headlines. It does not prompt parliamentary questions.

It appears years later, when organisations discover an ageing senior cohort with no properly trained generation behind them. When the knowledge pipeline has run dry and nobody quite knows when it happened.

The handloom weavers did not lose overnight. They lost gradually, and then all at once.

The hiring signal tells you we are still in the gradual phase. History tells us what tends to follow.

Part Two: The Bull Case

Altman’s Abundance

Sam Altman has become the most prominent voice for the optimistic interpretation of what AI will do to the economy. His argument deserves to be taken seriously, not dismissed as the self-serving cheerleading of a technology CEO with a financial interest in the outcome. The substance is more rigorous than the critics usually acknowledge.

His central argument is straightforward. The cost of intelligence is collapsing.

This is not a metaphor. It is an economic claim about the marginal cost of producing cognitive output. Legal drafts, financial models, computer code, strategic analysis, research summaries: the cost of generating each of these is falling rapidly towards zero. Not to zero yet. But directionally, and accelerating.

History offers a clear guide to what happens when something scarce and expensive becomes cheap and abundant.

Everything changes. The products and services built on top of that newly cheap input become cheaper too. New products and services emerge that were previously uneconomic. The economic geography of entire industries is redrawn.

Altman’s most striking formulation is the one-person unicorn. Within this decade, he argues, individuals running networks of AI agents across coding, legal work, finance and marketing will build billion-dollar companies alone. The capital, expertise and headcount that once formed the barriers to building serious businesses will dissolve.

The idea sounds extraordinary. It no longer looks obviously wrong. More about that next week.

Inside Yorkshire AI Labs, I routinely see early-stage founders compressing work that previously required entire teams into the capacity of a single individual augmented by AI. Founders conducting strategic software development, marketing communications, technical architecture and market analysis simultaneously, using AI for the information-processing components of each. The output per person has changed materially. The change is visible week to week.

The academic underpinning for the optimistic view comes primarily from Erik Brynjolfsson at Stanford. His productivity J-curve thesis argues that the economic impact of general-purpose technologies typically arrives with a significant lag. Firms adopt the tools before they have redesigned the processes required to get the full benefit from them. The productivity gains appear at the individual level well before they appear in aggregate economic statistics.

This is not an ad hoc defence of AI. It is a documented historical pattern.

Electrification began transforming American factories in the late nineteenth century. The first electric motors appeared in factories in the 1880s. The productivity gains from electrification did not appear clearly in macroeconomic data until the 1920s. Forty years after the technology was deployed.

The delay was not because electrification failed. It was because getting the full benefit required factory owners to fundamentally reorganise their buildings, their workflows and their management structures around the new technology rather than simply inserting electric motors into factories designed for steam power. That reorganisation took decades.

Once it occurred, the productivity surge was rapid and substantial.

There is a second argument that strengthens the bull case, drawn from the history of every previous technological disruption. Economists call it the reinstatement effect.

Every major technological revolution has displaced certain categories of labour. The steam engine displaced handloom weavers. The combustion engine displaced horse-related trades. Computers displaced large numbers of clerical workers. Each displacement was real, and the pain of the transition for those affected was also real.

But in every case, the technology simultaneously created entirely new categories of work that had not previously existed at scale. Railways eliminated traditional transport work but created railway engineers, track maintenance crews, station staff, logistics planners, freight insurers and a new commercial geography that generated employment across every sector it connected.

The optimists argue that AI will follow the same pattern. The new jobs it creates may not yet have names. Nobody in 1995 would have predicted the emergence of social media managers, platform trust and safety officers, user experience researchers or data protection consultants. Those roles did not exist yet because the platforms that would require them had not been built.

The bull case rests on the argument that AI will create new demand that neither Anthropic’s researchers nor anyone else can yet see.

That argument has been directionally correct across every previous major technological transition. The industrial revolution, electrification, computerisation, the internet: each produced dislocation, and each eventually produced substantially more employment than it destroyed.

The question is whether this transition is fundamentally different in a way that breaks the pattern.

Part Three: The Bear Case

The Great Decoupling

The pessimistic interpretation of AI’s economic impact does not come from technophobes or reactionaries. Some of the most rigorous economists working today hold this position, and their argument deserves the same serious treatment as the bull case.

Their argument rests on a simple observation that carries significant weight.

Every previous technological revolution disrupted physical labour. The steam engine, the power loom, the internal combustion engine, the industrial robot: all of these replaced human muscle. They were immensely consequential, and they caused genuine hardship for those whose livelihoods depended on the physical tasks they displaced.

But they did not replace the human mind.

For two centuries the economic value of educated labour rested on a foundation that technology could not touch. The wage floor for cognitive work was sustained by the fact that machines could not do cognitive work. The twentieth-century professional middle class was built, in its deepest economic logic, on the gap between what machines could do with matter and energy and what only human minds could do with information and judgment.

AI closes that gap.

Daron Acemoglu at MIT describes the most important risk as so-so automation. His concern is not that AI fails to generate productivity gains. It is that firms deploy AI primarily to reduce labour costs rather than to create genuinely new economic value.

In that scenario you replace the paralegal with a document review system. You do not develop a new legal product that uses the freed capacity to serve clients who previously could not afford legal services. You do not create new forms of legal work requiring different skills. You simply bank the margin.

Corporate productivity statistics improve. Margins expand. GDP figures remain healthy.

But wages for the majority of knowledge workers stagnate because the supply of competent output has risen relative to demand for human-generated work. Growth continues, but prosperity becomes decoupled from it. The aggregate numbers look fine while the distributional reality deteriorates.

This is not a theoretical risk. It is the pattern that characterised the management of deindustrialisation in large parts of the United Kingdom and the United States over the past four decades. Aggregate GDP grew. Technology improved. The communities where the displaced workers lived did not recover.

The Anthropic data contains evidence that points towards this risk rather than away from it.

The workers most exposed to AI displacement are not factory workers. They are not the low-skilled workers that the standard disruption narrative has always pointed at. The Anthropic study found that the most exposed workers are four times more likely to hold a graduate degree than the least exposed. They earn substantially more. They are 15.5 percentage points more likely to be female.

That last point deserves to be stated plainly rather than buried in a footnote.

The professionals most at risk from AI displacement right now are the educated women who built careers in legal, financial and analytical services over the past three decades. Not because they made different choices or lack relevant skills. Because the work those careers are built on happens to be the work that AI can currently do most effectively.

Look at the law firm. Look at the consultancy. Look at the investment bank. Look at the accounts department.

These are the workplaces where observed AI deployment is currently highest. And these are disproportionately the professional environments where women who fought to enter the knowledge economy over the past generation now work. The social and political implications of that distribution are considerable and are receiving almost no attention in the mainstream commentary.

Yuval Noah Harari pushes the argument into territory that economists find uncomfortable but which cannot responsibly be avoided.

Even if the distributional problem can be managed through redistribution, he argues, even if universal basic income or equivalent policies can preserve material living standards for those displaced, a deeper question remains.

What happens to societies where large numbers of people cannot find meaningful work?

Employment is not merely a source of income. It provides identity, purpose, daily structure and social status. The connection between meaningful work and individual psychological health is not speculative. It is documented across decades of research. The long term decline in male labour force participation in parts of the United States since the 1970s, and the well-evidenced correlation between that decline and rising rates of addiction, family breakdown and political extremism, tells you that the consequences extend far beyond economics.

Even if we solve the income problem, we have not solved the meaning problem.

If the professional middle class loses its economic function at scale, the political consequences will not wait for academic consensus to form. They will arrive, as they always have, in the form of disruption that surprises everyone who was watching the wrong indicators.

Part Four: My View

Intelligence Has Become Cheap

In my Yorkshire Post column this week I returned to a memory from my childhood that sits at the centre of everything this debate is actually about.

In the early 1970s my parents rented our colour television. That was entirely normal at the time. A good television cost around £250, which in today’s money represents well over £4,500. Families did not buy televisions. They entered a financial arrangement with a rental company, paid weekly, and hoped the thing kept working.

Three channels. Heavy buttons. An aerial that needed adjusting depending on the weather. A large piece of furniture that cost the equivalent of several months of ordinary wages.

Fast forward fifty years.

Today you can walk into almost any retailer and buy a vastly superior 4K flat screen for roughly the same nominal £250. Fifty years of inflation have passed and yet the sticker price has barely moved. The real cost has fallen by around 95 per cent. The worker on the National Living Wage can earn that television in approximately twenty four working hours today.

Now ask a different question.

Can the same worker easily afford a house?

No. The average semi-detached house in England now costs the equivalent of roughly nine years of median full-time earnings. That ratio has roughly doubled since the 1970s.

Housing has not experienced the same productivity revolution as televisions. Its price is shaped by planning systems, land constraints and financial structures that protect existing asset values rather than manufacturing innovation and global competition. The forces that made televisions cheap were deliberately allowed to operate. The forces that would make housing affordable were deliberately constrained.

This distinction is not peripheral to the AI debate. It is the centre of it.

AI is about to do to intelligence what automation did to television manufacturing. The marginal cost of legal analysis, financial modelling, coding, research, strategic advice and medical diagnostics is falling rapidly. Not to zero immediately. Not without friction. But directionally, and with increasing speed.

The question is whether the outcome resembles the television or the house.

One path resembles the television outcome. Intelligence becomes abundant and widely accessible. The founder in Sheffield who previously could not afford the professional services available to a corporation in London can now access legal drafting, financial modelling, regulatory guidance and market analysis at near-zero marginal cost. Small businesses compete with multinationals on information-intensive tasks for the first time. Productivity gains spread across the economy rather than concentrating in the firms large enough to afford the humans who previously provided those services.

If that abundance materialises, and the bull case says it will, then the productivity gains will be real, large and measurable. GDP per capita will rise. Corporate margins will expand. The economy will generate more output with fewer human hours of labour.

That is precisely the moment when the argument for universal basic income moves from aspiration to arithmetic.

The logic is straightforward. If AI generates a genuine productivity surplus, if the economy produces substantially more with substantially less human labour, then the resources to fund a serious universal income exist. Not a symbolic gesture. Not a pilot scheme for a handful of researchers to study. A universal payment set at a level that genuinely sustains a dignified life, funded by the productivity gains that AI makes possible.

This is not a radical position. It is a conservative one. Every serious analysis of technological displacement eventually arrives at the same conclusion: when the gains from automation are large enough and concentrated enough, redistribution is not charity. It is the mechanism that prevents the abundance from becoming a crisis.

The Luddites were not wrong about what was happening to them. They were wrong about how to stop it. The correct response to a productivity revolution that eliminates skilled work is not to smash the looms. It is to ensure that the wealth the looms generate does not flow exclusively to the people who own them.

We have been here before. The post-war settlement built the NHS, the welfare state and universal education on exactly this logic. The productivity gains of industrial capitalism were real. The question was who would share them. A generation of politicians decided the answer should be: everyone.

That decision requires making again. Now. Before the sorting is complete.

In this scenario we also invest in the wider social infrastructure a genuine transition requires. Retraining programmes that are funded, serious and designed for adults with twenty years of professional experience rather than for teenagers learning to code. A regulatory framework that requires AI productivity gains to carry some obligation to the workforce they displace, rather than flowing entirely to shareholders. And a universal basic income that treats the abundance AI creates as a shared inheritance rather than a private windfall.

The other path resembles the housing outcome. The efficiency gains flow primarily to capital. Firms capture the cost savings as margin without creating new economic activity. The educated professional class faces a sustained income squeeze as the supply of competent output rises relative to demand for human-generated work. Young people find the entry level rungs pulled up ahead of them. The political consequences are not unpredictable. They are already visible in countries where economic transitions were managed badly.

History is not reassuring about which path societies tend to take without deliberate intervention.

The social media era provides the most recent and instructive parallel. The first iPhone launched in 2007. The early academic evidence of algorithmic harm to adolescent mental health began appearing around 2014. Serious regulatory response in the United Kingdom and European Union arrived around 2022. Fifteen years between the deployment of the technology and any serious attempt to govern its consequences.

By the time regulation arrived, two generations of young people had been exposed to systems specifically optimised to maximise engagement regardless of psychological cost. The damage was already done. The walls arrived after the flood.

Anthropic’s researchers themselves offer a different historical comparison in the report. The labour market impact of AI, they suggest, may resemble the China trade shock more than it resembles COVID. The economic disruption from China’s entry into global manufacturing was visible in employment data from roughly 2003 onwards. The political consequences, measured in the collapse of consensus around open trade, the rise of economic nationalism and the reshaping of electoral politics across the western world, did not arrive fully until 2016.

A thirteen year gap between the economic signal and the political reckoning.

If that analogy holds, the 14 per cent decline in youth hiring in AI-exposed occupations that we see in today’s data is the 2003 signal. The political reckoning arrives in the early 2030s.

And we are currently doing approximately as much to prepare for it as western governments did to manage the China trade shock. Which is to say: very little.

By the time the aggregate statistics confirm the structural break, an entire generation will already have been sorted into winners and losers. The sorting is happening now. It is quiet, distributed and politically invisible, which is precisely what makes it dangerous.

The bull case is coherent. The bear case is credible. My honest view is that the outcome depends entirely on decisions being made right now, in boardrooms and in Parliament, by people who are largely focused on the next quarter and the next election.

Both time horizons are too short for the problem in front of them.

That is not a prediction. It is a warning.

Final Thought 🚀

The television my parents rented in the 1970s became cheap because we built the systems required to make it cheap. Manufacturing investment, global supply chains, technical standards, competitive markets and deliberate policy choices all played a role. None of it happened automatically. It was the accumulated result of choices made by governments, businesses and consumers over decades.

Intelligence is becoming cheap for similar reasons. Enormous research investment, computing infrastructure, data access and strategic deployment decisions are pushing the cost of cognitive work steadily downward. Those decisions were made by specific people in specific organisations with specific interests. The outcome we are heading towards is not natural or inevitable. It is the product of choices that were, and continue to be, made by human beings.

The technology itself will do what it does.

What is not predetermined is who benefits from it, who bears the cost of the transition, and whether the societies on the receiving end of this shift have the political will to manage it in a way that does not repeat every previous failure.

The question is not whether intelligence will become cheap. It will.

The question is not whether this will displace professional work. It will.

The question is whether we will make the choices required to ensure that what is coming builds a floor rather than widens a gap.

History suggests we will move too slowly.

The 14 per cent drop in youth hiring in AI-exposed roles is already in the data.

The question is whether anyone in power is watching.

Until next Sunday,

David

David Richards MBE is a technology entrepreneur, educator, and commentator with over 25 years in technology. He writes for the Yorkshire Post and publishes The Sunday Signal weekly at newsletter.djr.ai

The Sunday Signal | newsletter.djr.ai | Issue #44 | Sunday, 8 March 2026

Bottom Line Up Front

Last week’s issue, The Loom, the Layoff, and the Life Raft, made the historical case. From Yorkshire’s handloom weavers to the junior developers of 2026, the pattern was clear. Skilled work gets eaten by machines. The people in the middle convince themselves it is temporary.

The response in my inbox was predictable. Developers telling me AI hallucinates. Lawyers explaining it still needs supervision. Professionals arguing it is not production-ready.

That may all be true.

But when I want to know whether something is real, I do not start with opinions. I start with markets.

Commentary is frictionless. Capital is not.

Financial markets are where pension funds, sovereign wealth, boards and institutions reprice assumptions with real money at stake. When billions are marked down in a single session, something structural has shifted.

This week, three separate corporate stories told the same story.

AI has stopped being a productivity narrative. It is becoming a cost narrative.

And the market is confirming it.

PODCAST: Listen To This Edition

Part One: The Triple Warning Shot

IBM and the fragility of “complexity”

On Monday 23 February 2026, IBM’s shares fell roughly 13% in a single trading session. Its worst single-session fall in over twenty-five years.

That is not a routine wobble. That is a repricing event.

IBM’s modern business is built heavily on consulting and legacy modernisation. For decades, the complexity of COBOL estates inside banks and governments acted as a moat. Untangling forty years of accumulated business logic required armies of specialists. That slowness was billable.

Then Anthropic published analysis suggesting tools like Claude Code can automate significant portions of COBOL exploration and reverse engineering.

Whether that claim proves fully accurate is secondary. The market heard something simpler.

Complexity might no longer be scarce.

If the moat weakens, the margin compresses. Public markets do not wait for certainty. They discount direction.

Thirteen per cent in a day is what direction being repriced looks like.

Atlassian and the end of the per-seat assumption

Over the past fifty-two weeks, Atlassian has been widely reported as down roughly 73%, trading near levels last seen in 2018.

Strip away macro noise and a cleaner question emerges.

Atlassian’s model is built on seats. Seats track headcount. Headcount is the input.

If AI increases output per engineer materially, companies need fewer engineers. If they need fewer engineers, they need fewer seats.

When input shrinks structurally, revenue assumptions shift structurally.

This is not about Jira features. It is about labour intensity. And it matters beyond Atlassian. There is an entire category of SaaS businesses whose revenue assumptions are built on human headcount at the companies they serve. The per-seat pricing model that defined a generation of cloud businesses is being repriced in real time. The transition will be brutal for the companies caught in between.

Block and the incentive structure

On Thursday, Jack Dorsey announced that Block would cut over 4,000 employees, roughly 40% of its workforce, explicitly linking the restructure to intelligence tools. The company is profitable. It is growing. No crisis. No failed product.

Shares jumped roughly 25% in after-hours trading.

That is not sentiment. That is capital signalling approval.

Every board in the developed world just saw a case study: reduce headcount aggressively, attribute it to AI efficiency, get rewarded.

The copycat cycle has begun.

Part Two: This Is How We Now Build at Yorkshire AI Labs

Let me be precise.

This is not speculation about the future of software. This is how we now model and capitalise companies inside the Yorkshire AI Labs portfolio.

Three years ago, a new software venture followed a predictable structure. Core development squad of six. UI and front-end another six. QA and testing another six. DevOps and release engineering another six. Twenty engineers before meaningful traction was considered normal. Investors expected it. Headcount signalled seriousness.

Now we increasingly structure around one highly capable AI-native engineer at the centre of the build.

One.

Not because we are cutting corners. Because the economics have changed.

Even in Yorkshire, a fully loaded engineer costs roughly £80,000 to £100,000 per year once salary, national insurance, pension, tooling and overhead are included. Twenty engineers means £1.6 million to £2.5 million annual burn before product-market fit. Under our current portfolio model, a single AI-native engineer, properly tooled, can orchestrate generation, testing, documentation, deployment and integration work that previously required multiple squads.

The cost base collapses. Runway extends. Dilution reduces. Capital efficiency improves dramatically.

That is not theoretical. It is how we are allocating capital today.

A Live Portfolio Example

A CTO in one of our portfolio companies recently needed to replicate a major bank integration environment. APIs. Error handling. Rate limiting. Latency behaviour.

Historically, that would have required weeks of structured engineering work.

Using AI tooling, he generated a functioning mock banking environment in minutes. He reviewed it, validated it and moved straight to live integration.

Four weeks became minutes - literally.

When you see that inside your own portfolio, you understand why public markets are moving.

The Pattern Across Our Portfolio

The banking example is one instance. The pattern repeats.

A fintech building KYC compliance tooling used to need a legal team, a compliance team and an engineering team working in parallel. Now one AI-augmented compliance engineer produces a first-draft mapping from regulation text to technical specification in hours. The parallel-squad model has collapsed into a sequential, much leaner process.

A health-tech company building NHS data integrations used to require specialists in HL7 FHIR standards who took years to develop that specialism. Now AI surfaces the relevant standards, generates the integration schemas and flags known edge cases, compressing the learning curve dramatically.

A B2B SaaS company building workflow automation used to need separate product, research and engineering teams. Now, a two-person team does the work that previously required ten.

This is what last week’s issue called the Slow Squeeze. Wages under pressure first. Status following. The profession reorganising around fewer people. This week’s three market signals are what the Slow Squeeze looks like when it hits a public market all at once.

Part Three: The K-Shape. Which Side Are You On?

Understanding what is happening at the macro level is one thing. The more important question is what it means for the individual.

Miles Deutscher has been writing about what he calls the K-shaped AI economy: one diverging trajectory where those who master AI compound their advantage, and another where those who do not are gradually managed out. A visualisation circulating widely this week, a dot matrix attributed to Damian Player, CEO of Agent Integrator, puts the adoption split into stark relief:

Each dot represents 3.2 million people. The full grid is 8.1 billion humans alive today.

Of those 8.1 billion, approximately 84% have never meaningfully used AI. The revolution, for most of the planet, is still entirely abstract. Another 16% are free chatbot users: occasional, modest, competitive advantage thin and eroding. Approximately 0.3%, some 15 to 25 million people, pay for AI tools and are gaining genuine efficiency advantages. And at the absolute frontier, roughly 2 to 5 million people, 0.04% of humanity, are using coding scaffolds to build: deploying tools like Cursor, Bolt, or Claude Code to create their own applications, automations and infrastructure.

That last group is the Red Dot. The gap between the Red Dot and everyone else is widening at an accelerating rate.

The Four Baskets

Deutscher maps individual AI capability into four levels.

Basket One: The Basics. Using AI for summaries, research, draft emails. Most people who use AI at all are here. It saves time. But as a competitive differentiator, it is rapidly approaching zero. If everyone can summarise a document with a chatbot, doing it with a chatbot confers no advantage. These skills are being commoditised in real time.

Basket Two: The Power User. Advanced prompting. Multi-step agentic workflows. Automating sequences of work without manual intervention at each step. At this level, you genuinely do the work of two or three people. The efficiency gain is real. But the uncomfortable truth: you are still an employee, still trading time for money, still vulnerable to the next efficiency leap that makes your current toolkit look basic.

Basket Three: The Builder. This is where things change qualitatively. You are using coding scaffolds to create your own applications, your own automations, your own solutions to your own problems. You stop being a consumer of other people’s software and start being a producer. You build tools that did not previously exist, in hours rather than weeks. You are no longer dependent on the off-the-shelf tool market. You create bespoke capability at low cost.

Basket Four: The Master. AI-native businesses. Consulting. Distribution. The connection between time spent and income earned is fully broken. You are building systems that operate without you. The IBM signal, viewed from Basket Four, is not a threat. It is an opportunity. The COBOL modernisation work that IBM charged armies of consultants to do is now available to a single AI-native practitioner with the right tools and the right clients.

Most professionals reading this are in Baskets One or Two. The direction of travel matters more than the current position. The window to move up is narrowing. Not closing. Narrowing.

The Ladder Is Being Pulled Up

The entry-level roles that once trained professionals, the junior analyst, the graduate developer, the trainee paralegal, are the first to go. Not because AI does their work perfectly. Because AI does it well enough, cheaply enough, quickly enough that companies are no longer willing to pay a human to do it while they learn.

Professional development has always worked through apprenticeship. You start at the bottom. You do the grunt work. You accumulate the judgment and contextual knowledge that makes you valuable at the senior level over the years. The junior role was not just a job. It was the training ground for the senior role.

If the junior role disappears, the training ground disappears with it. The people who climbed that ladder got something that will not be available to the next generation. The ladder is being pulled up behind them.

Last week’s issue made the point from history: the handloom weavers also had genuine expertise. The new categories of work in the textile industry were not available to them. The transition, though not instantaneous, was fast enough to destroy a generation.

Deutscher’s argument is that we are in a 6 to 12-month window where the gap between the two trajectories of the K is still bridgeable. After that, the compounding advantages of the Red Dot become structurally difficult to close.

The timeline is debatable. The direction is not.

What You Can Actually Do

The analysis is only useful if it leads somewhere actionable.

The first move is the mindset shift from passive observer to active experimenter. Most people in Baskets One and Two are using AI the way people used early search engines: to find information faster, not to build new capability. That framing limits what you can learn and what you can build.

The second move is systematising your exposure. The people getting ahead are not occasional readers of good AI articles. They have a daily or weekly discipline around understanding what is now possible: research announcements, developer communities, product releases that signal new capabilities. Not to become technical experts. To understand what tools now exist before those tools become mainstream knowledge.

The third move, the one that separates Basket Two from Basket Three, is to build something. Anything. A custom automation for a task you do repeatedly. A tool specific to your work. A dashboard that aggregates information you currently gather manually. The act of building with AI tools transforms your understanding of what is possible. It moves you from consumer to producer, which is exactly the distinction the market is beginning to price.

The fourth move is positioning. The Red Dot individuals are connected to other builders, other AI-native practitioners, other people using these tools seriously. The networks forming around AI-native practice are becoming economically significant. The people inside them are getting access to opportunities and knowledge not visible from outside.

The IBM signal, the Atlassian signal, and the Block signal are all pointing in the same direction: companies are restructuring around fewer, higher-capability humans deeply integrated with AI. The people who survive and thrive are the ones who made themselves irreplaceable by becoming builders rather than users.

Final Thought 🚀

Last week set out the historical pattern. This week, the market confirmed it.

IBM’s repricing. Atlassian’s long decline. Block’s headcount reduction rewarded by investors.

Three separate stories. One structural signal.

When the cost of producing digital output falls, labour intensity falls. When labour intensity falls, margins shift. When margins shift, valuations move.

The machine did not need to be perfect. It needed to be cheaper and capable of improvement. By the time it closed the quality gap, the industry had been rebuilt around it, with far fewer people, different skills, and no apology for the transition.

Opinions are frictionless.

Markets are not.

Capital has started to move.

The window is open. For now.

Until next Sunday, David

David Richards MBE is a technology entrepreneur, educator, and commentator with over 25 years in technology. He writes for the Yorkshire Post and publishes The Sunday Signal weekly at newsletter.djr.ai

The Bottom Line Up Front

Two hundred years ago, Britain’s most skilled craftsmen watched their livelihoods unravel. Not because they lacked talent. Because the machine did not need their talent. It needed their output.

The software developers, paralegals, and junior analysts of 2026 are living through the same story. The machinery is different. The economic logic is identical.

This week: the historical pattern that explains everything about the AI jobs crisis, the ten roles at the sharpest end of the displacement curve, what my latest Yorkshire Post column reveals about who actually wins, and a concrete three-phase plan for anyone who wants to own the machine rather than be replaced by it.

This issue is also available as a podcast — an experiment using an AI-generated version of my voice. Listen on Spotify or Apple Podcasts and tell me what you think.

Part One: The Weaver’s Warning

The Canary That Nobody Listened To

There is a story from the English Midlands that every professional facing AI displacement should know by heart. It begins around 1810. It ends in migration, poverty, and the permanent extinction of a skilled trade. And it contains every lesson we need for 2026.

The handloom weaver was the software developer of the early nineteenth century.

This was not a low-skill labourer. This was a craftsman with years of apprenticeship behind them, a sophisticated understanding of materials and mechanics, a home-based operation with genuine autonomy, and a wage that placed them firmly in what we would today call the middle class. In the late 1700s, a skilled weaver in Lancashire or Yorkshire could earn enough in three or four days to support a family for the whole week. They owned their tools. They set their own hours. They had leverage.

That leverage came from knowledge you could not simply acquire. The craft required mastering the tension of the threads, the rhythm of the shuttle, and the relationship between warp and weft. This knowledge was the moat. It was the thing that made a weaver irreplaceable.

Until, of course, it wasn’t.

The Slow Squeeze (1815–1830s)

The power loom did not arrive like a cannonball. It arrived like a tide. Gradual, inevitable, and ultimately overwhelming.

Between 1815 and the 1830s, weavers did not lose their jobs overnight. Something more insidious happened. The price of their output collapsed. A steam-powered loom could produce twenty times more cloth per hour than a human. Factory owners did not need to fire the weavers. They simply drove the market price of cloth so low that weavers had to work twice as hard for half the money just to survive.

The weekly wage fell from roughly 25 shillings in 1800 to just 5 shillings by 1830. The same hands. The same skill. A fifth of the income.

They argued that machine-made cloth was inferior. For a while, they were right. But the machines improved. Then the consumers stopped caring about the difference.

Sound familiar?

In 2026, a junior copywriter argues that AI-generated content lacks soul. A Tier-1 support agent argues that customers prefer talking to a human. A junior developer argues that AI code requires too much review to be reliable. They are not wrong. Not yet. But the machines are improving. And the economics are already shifting.

This is the Slow Squeeze. The job title survives. The wage, the job security, and the career trajectory quietly erode.

The Luddite Rebellion: The Emotional Breaking Point

As their livelihoods collapsed, the weavers did not simply retrain. They fought back. Organised groups of skilled textile workers broke into factories at night to smash the power looms with sledgehammers.

The popular myth is that the Luddites were ignorant people who feared technology. That is a comfortable story for those who benefit from automation. The reality is more complex.

The Luddites were not anti-technology. They were anti-exploitation.

The machines were being used specifically to bypass existing labour laws, undercut agreed wage structures, and destroy a standard of living that skilled workers had built over generations. Their grievance was not with the loom itself. It was with the fact that the productivity gains were going entirely to the factory owners.

Parliament made machine-breaking a capital offence, punishable by death. Soldiers were deployed. The movement was crushed. The weavers lost.

By the 1840s, the handloom weaver as a career was functionally extinct. Many ended up working in the very factories they had fought against, no longer as independent craftsmen but as machine overseers. The value had migrated from weaving the cloth to maintaining the loom. Those who could not make that transition emigrated, thousands of them, to the United States and Australia.

Their children benefited. Lower cloth prices meant cheaper clothing for everyone. New industries emerged. The weavers lived through the worst of it. Their grandchildren inherited the benefits.

This is the U-shaped recovery. And it is exactly the shape of the curve that today’s displaced professionals are at the beginning of.

1826 vs. 2026

The junior software developers, paralegals, market research analysts, and content writers of today are the handloom weavers of 1820. Talented. Genuinely skilled. The apprenticeship was real. The moat existed.

But the power loom has arrived.

The task is no longer to weave faster than the machine. That race is over. The task is to understand who owns the factory.

Part Two: The Ten Jobs in the Crosshairs

As we move through 2026, the conversation has stopped being theoretical. AI is no longer the future of work. It is a standard employee in offices across the world.

The research below draws on Goldman Sachs, the World Economic Forum’s Future of Jobs Report 2025, McKinsey’s State of AI 2025, Gartner’s Predicts 2026, the OECD Digital Education and Employment Outlook, and ADP’s latest labour trend data. This is not speculation. It is current employment data.

1. Customer Service Representatives (Tier 1)

Agentic AI has moved far beyond the frustrating chatbots of 2022. Autonomous agents can now process refunds, troubleshoot faults, and handle complaints with voice responses indistinguishable from a human. Tier-1 support is almost entirely automated at the companies moving fastest. Gartner’s analysis notes that total replacement leads to a measurable drop in customer sentiment, so human agents survive but only in conflict resolution. The complex complaint. The edge case the AI cannot navigate without reputational damage. The weaver is now the loom overseer.

2. Data Entry and Document Clerks

AI vision and OCR have reached a precision in 2026 that makes manual data entry indefensible. The WEF’s Future of Jobs Report confirms clerical and administrative roles are among the fastest-declining categories between now and 2030. The role transforms into Data Quality Auditor. The skill required is not typing speed. It is pattern recognition and judgment.

3. Junior Software Developers

For two decades, learning to code was the single most reliable piece of career advice a young person in Britain could receive. McKinsey’s State of AI 2025 identifies coding as one of the functions seeing the fastest acceleration in agentic AI deployment. The entry-level runway is gone. New developers entering the market are expected to be System Architects and AI Orchestrators from day one. More on this in Part Three.

4. Paralegals and Legal Researchers

Specialised Legal LLMs can review thousands of pages of discovery documents, summarise case law, and flag inconsistencies in a fraction of the time it takes a junior associate. Goldman Sachs specifically identifies legal assistants as among the roles with the highest exposure to generative AI. Law firms are not waiting. Junior headcount is being reduced. The value moved from reading the brief to arguing the case.

5. Bookkeepers and Payroll Specialists

Accounting software in 2026 is effectively self-reconciling. AI automatically categorises expenses, detects anomalies, reconciles bank feeds in real time, and handles routine tax filing. ADP’s labour data shows Professional and Business Services experiencing a measurable cooling in job growth, attributed directly to AI-driven efficiency gains. Strategic Tax Planning and complex financial advisory are where the human value now lives.

6. Basic Content Writers and SEO Specialists

Short-form content, product descriptions, and SEO-driven articles are now generated at scale by AI. In 2026, search engines weight Information Gain, the degree to which content adds something genuinely new. Generic content no longer ranks. The shift is from writer to Content Strategist and AI Editor: the person who knows what angle to take and what the AI confidently gets wrong. As the WEF notes, creative thinking is rising in importance precisely as execution becomes automated.

7. Junior Market Research Analysts

AI can scrape the web, conduct sentiment analysis, and generate a fifty-page market report in under a minute. Gathering data is no longer a skilled role. Interpreting what that data means for a CEO’s five-year strategy, identifying the pattern the data does not show, that is where the analyst earns their salary. McKinsey confirms knowledge management is now one of the leading areas for agentic AI scaling.

8. Telemarketers and Outbound Sales

Cold calling has been taken over by AI voice agents running thousands of simultaneous conversations, never experiencing call reluctance, improving in real time. Goldman Sachs lists telemarketers among the highest-risk occupations for AI displacement. The human sales professional survives only at the warm end of the funnel. When the relationship requires depth. When the negotiation requires reading the room. The cold call is not.

9. Proofreaders and Translators

Neural translation has reached near-perfect functional fluency for most major world languages. Goldman Sachs lists proofreaders and copy editors among the most exposed roles. Survival territory is narrow but real: high-stakes literary translation, legal contracts where precise human accountability is a regulatory requirement. The volume of routine work that once kept professional linguists employed is gone.

10. Template-Based Graphic Designers

For small businesses, the requirement to hire a designer for a logo, a social media asset, or a basic website layout has effectively disappeared. What survives is Creative Direction: knowing what good looks like, what a brand needs to communicate, and what the machine is producing that is technically proficient but strategically empty. The WEF’s fastest-declining jobs data confirms graphic and multimedia designers are under significant structural pressure.

Part Three: Be Nice to the Nerds. While You Still Can.

Building on this week’s Yorkshire Post column

I used to give young people very clear advice. Learn to code. It is the key. I believed it so completely that I helped open a coding academy.

This week, that advice required significant revision.

The End of Code as Permission

To understand what is happening, you have to understand what coding actually was during the Internet era. It was not just a technical skill. It was permission.

Creative people had ideas. They needed a translator who spoke machine. The designer could imagine the product but could not build it. The entrepreneur could see the market but could not ship the software. They all hit the same wall: you can imagine it, but can you build it?

That asymmetry gave the programmer their power. This is the story of the Winklevoss twins and Divya Narendra, who conceived a social network called HarvardConnection and needed someone to build it. They recruited Mark Zuckerberg. The creative spark was not the scarce resource. The ability to execute in code was. The translator had leverage over the person with the idea.

That dynamic has now changed.

AI can produce working software of startling quality. What once required a team of engineers working for weeks can increasingly be achieved by a single person with a well-crafted prompt and sound judgment. The machine speaks fluent software. The translator is no longer the bottleneck.

What This Actually Means

If AI can write the code, learning to code is no longer a moat. It becomes table stakes. The equivalent of knowing how to use a spreadsheet. Necessary, perhaps. Sufficient, no longer.

Dario Amodei, Chief Executive of Anthropic, told The Economist at Davos that AI could handle most, possibly all, of what software engineers currently do within six to twelve months. His own team confirmed that at Anthropic, the code is now effectively one hundred per cent AI-generated. Elon Musk has gone further, predicting traditional coding will be obsolete by the end of this year.

I agree with the direction, if not every detail. The mass-market demand for armies of developers writing routine business software is the part most exposed to replacement. The OECD has already identified a significant cooling in entry-level white-collar hiring across member countries through late 2025 and into 2026.

The Upside, and It Is Significant

When execution is democratised, originality becomes the new scarcity.

Domain experts who understand real problems can now build tools without waiting for a technical co-founder. The Winklevoss problem is solved. You no longer need Zuckerberg to build the thing. You are Zuckerberg now.

The winners of the next decade will not be those who can write a function in Python. They will be the people who think clearly, see problems others miss, understand the domain deeply enough to know what quality looks like, and can direct increasingly capable machines toward genuinely useful outcomes.

Be nice to nerds, Gates said. The advice still holds. It is simply that the nerds of tomorrow may look nothing like the nerds of today. They may be manufacturers who understand industrial processes. Lawyers who understand regulatory complexity. Teachers who understand how children actually learn.

The handloom weaver’s skill was weaving. That skill became worthless. But the knowledge of textiles, of markets, of quality survived and migrated into new roles. The lesson is not to abandon what you know. It is to stop confusing the tool you use with the knowledge you hold.

Part Four: How to Be AI-Proof

The 90-Day Survival Plan

The weavers who survived the power loom were not the ones who wove fastest or protested loudest. They were the ones who understood what the machine could not do, positioned themselves where human judgment was irreplaceable, and moved quickly enough to get there before everyone else did.

Three phases. Ninety days each.

Phase 1: AI Literacy (Days 1–90) Focus: Not coding. Logic.

Real AI literacy means understanding how Agentic AI reasons, where it excels, and where it fails with complete confidence. Where it hallucinates. What kinds of tasks it reduces to triviality and what kinds remain genuinely hard. The WEF identifies AI and big data literacy as the fastest-growing skill category in the 2025–2030 period.

Milestone: Conduct a full AI Audit of your own job.

List every significant task you do in a working week. Sort them into two columns.

Column A: tasks that take 50% or more of your time but require none of your unique knowledge, judgment, or relationships.

Column B: tasks where your specific experience, relationships, or accountability are genuinely what make them valuable.

Column A is your displacement risk. Column B is your survival territory. If Column A is longer than Column B, you need to start moving.

Phase 2: Orchestration (Days 91–180) Focus: Multi-tool integration.

The value in 2026 is not in knowing the tools exist. It is in being able to string three or four AI agents together into a workflow that replaces a process that currently takes human time. This is the difference between using a tool and being a system architect.

Milestone: Build a Digital Twin of your current workflow.

Try Zapier, n8n, or Claude Code to automate the tasks from Column A. The goal is not perfection. The goal is to do it, to understand how agents pass information between each other, where the failure points are, and where a human still needs to make a judgment call.

Once you have built even a basic automated workflow, you have crossed a line. You understand how the machine thinks because you have made it think.

Phase 3: Specialisation (Days 181–270) Focus: Domain-specific logic.

The most AI-resistant professionals in the world share one characteristic: they hold knowledge that is genuinely hard to acquire, hard to encode, and consequential when it goes wrong.

AI can summarise what is known. It cannot replace the person who has spent twenty years learning what is not written down.

Milestone: Get certified in AI Governance or Strategic Risk Management within your specific field.

The Alan Turing Institute, BCS (Chartered Institute for IT), and MIT Sloan Executive Education all offer relevant programmes. Domain depth plus AI fluency plus accountability is the new moat. It is considerably harder to replicate than the ability to write a loop in Python ever was.

The Honest Assessment

None of this is comfortable. The weavers who thrived after the power loom arrived were a minority. Most of that generation struggled. The U-shaped recovery is real, but the bottom of the U is a brutal place to spend a decade.

The difference between 1826 and 2026 is the speed of the information available. The weavers of Lancashire did not have access to a weekly newsletter explaining the pattern they were inside. They did not have McKinsey telling them their core skill was being automated. They did not have the WEF telling them at Davos that their entire career category was being structurally reorganised.

You do.

The question is what you do with it.

🚀 Final Thought

The power loom did not destroy weaving. It destroyed weavers who refused to understand that their value was never in the motion of the shuttle. It was in the knowledge of the cloth.

The AI revolution will not destroy your career. It will destroy the version of your career that requires you to move the shuttle.

The question is not whether to adapt. It is whether you adapt before the wage spiral begins, or after.

The weavers who made it understood one thing early: the machine is not the enemy. The person who already owns the machine and has replaced you is.

Don’t be replaced. Own the machine.

Until next Sunday,

David

David Richards MBE is a technology entrepreneur, educator, and commentator with over 25 years in technology. He writes for the Yorkshire Post and publishes The Sunday Signal weekly at newsletter.djr.ai

Boards love the word “professionalise.”

It sounds responsible. Mature. Sensible.

In practice, it often means removing the person who made the company worth owning in the first place.

This week’s Yorkshire Post column was not theory. It was memory. I was fired in a hotel in Belgravia, with a document slid across a table, by a chairman I had appointed. Then I was reinstated days later.

The point is not that founders are infallible.

The point is that founders are not interchangeable.

History shows what happens when boards forget that. Sometimes they destroy billions in value. Occasionally they save the business.

At the same time, the largest companies on earth are making infrastructure bets so vast they are beginning to bend markets and power grids around them. Not hundreds of billions. We are now looking at combined capital expenditure approaching seven hundred billion dollars in a single year.

Two decisions.

Who leads.

And who builds.

Both define the next decade.

Subscribe now

Raise Money. But Never Give Away the Keys.

Every ambitious founder should raise capital.

Whether that means venture capital, public markets or both, external money is how small companies become important ones. Capital gives you speed. Speed gives you market position. Market position gives you options.

The mistake is not raising money.

The mistake is giving away control before you have proven what you are building.

The pattern is painfully consistent. A founder with a strong idea and early traction takes a pre-seed or seed round. The term sheet looks generous. The valuation feels fair. But buried in the details are board seats, veto rights, anti-dilution clauses and preferred liquidation terms that shift power long before the company has found its footing.

By the time the founder realises what they signed, the investor is not a partner.

They are the landlord.

This is the difference between venture capital and private equity. Venture capital, in its best form, backs founders and accepts the wild variance that comes with early-stage conviction. Private equity buys control, installs management, optimises operations and extracts returns on a defined timeline.

Both models work. But they serve different purposes at different stages.

What is happening increasingly, and I see it locally, is venture investors behaving like private equity at the earliest possible stage. Taking board control at pre-seed. Installing their own chair before the product has reached market. Firing founders in the first innings of the game, not because the company is failing, but because the founder is not docile enough for the investor’s operating model.

This is not governance. It is capture.

When I was asked to sign that document in a hotel in Belgravia, I was a CEO who had built a company, taken it public and created hundreds of millions in value. And still the board I had constructed moved to remove me. I was reinstated days later. But the lesson stayed.

If it can happen to a founder at that stage, it can happen to anyone. And it is happening now to founders who have barely started.

A VC who removes a founder at seed stage is not protecting an investment. They are destroying the only thing that makes the investment worth making. At that point, the founder is not one part of the machine. They are the machine. The vision, the conviction, the customer relationships, the product instinct, the pace. Remove them and you are left with a cap table, a pitch deck and a set of Notion documents.

The lesson is not to avoid raising capital.

The lesson is to raise it wisely.

Understand what you are signing. Every term sheet is a transfer of power disguised as a transfer of money. Board composition, voting thresholds, reserved matters, drag-along rights. These are not admin. They are architecture. Get a lawyer who works for you, not one recommended by the investor.

Retain board control for as long as you can. At pre-seed and seed, there is no reason for a founder to surrender a board seat. If an investor demands one as a condition of a modest cheque, they are telling you exactly how they intend to behave when things get difficult.

Choose investors who have backed founders through adversity, not just through growth. Ask the question directly. Have you ever fired a founder? How early? Why? The answers will tell you everything.

Look at dual-class structures early. They are not just for IPOs. Zuckerberg controls Meta with roughly thirteen percent economic ownership because he holds the majority of voting shares. That structure was established long before the company went public. Founders who wait until listing to think about control have already left it too late.

Raise money. List when the time is right. Both are signs of ambition and confidence.

But never confuse a cheque for alignment.

And never assume that the person sitting across the table at your board meeting is building the same company you are.

When Firing the Founder Destroys Value

And When It Saves It

There is a pattern in corporate history. Boards replace visionary founders with operational leaders in pursuit of calm.

Calm is usually expensive.

Consider Apple after the departure of Steve Jobs in 1985.

Apple did not implode. It diffused.

The Newton absorbed huge resources and became a cultural punchline for its flawed handwriting recognition. The Pippin was a forgettable attempt at gaming hardware. The Macintosh TV felt like a product without a thesis. The desktop line multiplied into Quadras, Performas and Power Macintoshes with marginal differences and mounting confusion.

There was activity everywhere. But no clarity.

By 1997 Apple was widely described as being within ninety days of running out of cash. Market share had collapsed from around sixteen percent to less than four. Annual losses hit $1.03 billion. The company that invented the personal computer revolution was being written off.

The board acquired NeXT, effectively to bring Jobs back. He eliminated around seventy percent of the product line. He cut the bewildering range of desktops to one. He cut portables to one. He negotiated a $150 million investment from Microsoft, of all companies, to keep the lights on. Ruthless focus returned. So did profitability.

The difference was not engineering talent.

It was taste.

Taste is the ability to say no. Taste is strategic subtraction. Taste is conviction before consensus.

Now look at Yahoo and Jerry Yang.

In February 2008, Microsoft offered $44.6 billion to acquire Yahoo. The board, led by Yang, rejected the bid as “substantially undervaluing” the company. Microsoft raised its offer to $33 per share. Yang rejected that too. Shareholders revolted. Carl Icahn joined the board. Yang stepped down as CEO in November 2008. The company cycled through four more chief executives in five years, and eventually sold its core business to Verizon in 2017 for roughly $4.5 billion.

That is not poor execution. That is generational value destruction on a staggering scale.

Or take OpenAI removing Sam Altman in late 2023. Within days, over ninety-five per cent of employees signalled they would leave. Microsoft offered to hire the entire workforce. The board discovered a brutal truth. Governance without the founder meant having no company to govern.

Those are the disasters.

But firing a founder is not always wrong.

BlackBerry is a case study in founders misreading a platform shift. Co-CEOs Jim Balsillie and Mike Lazaridis clung to physical keyboards while Apple’s iPhone remade the entire category around touchscreens. After both stepped down in January 2012, the company pivoted into enterprise software and cybersecurity. It did not regain dominance, but it survived.

Tesla replaced its founding CEO Martin Eberhard in 2007 amid cost overruns and execution strain on the original Roadster programme. The leadership that followed was volatile and aggressive. It also forced an entire industry to electrify.

So what is the dividing line?

It is not charisma. It is not volatility.

It is whether the founder is still directionally right.

If the founder is still the engine, removal destroys value.

If the founder has become the brake, removal can unlock it.

Boards often fail because they remove founders for the wrong reason. They remove them because they are uncomfortable.

Professional operators optimise what exists. Founders create what does not.

If you replace the explorer too early, you institutionalise mediocrity.

The Seven Hundred Billion Dollar Gamble

While boards debate personalities, markets are engaged in a different act of conviction.

Amazon, Alphabet, Meta and Microsoft are collectively committing capital expenditure of between $635 billion and $700 billion in 2026. That is roughly double their combined spend in 2025, which was itself a record. A decade ago, these same four companies spent $31 billion between them. The scale of acceleration is without precedent in corporate history.

This is no longer asset-light software. This is industrial-scale infrastructure.

Amazon alone has guided $200 billion in capital expenditure this year. Alphabet is targeting $175 billion to $185 billion. Meta between $115 billion and $135 billion. Microsoft is tracking towards $120 billion. Even Oracle, which lacks the free cash flow of the hyperscalers, is raising $45 billion to $50 billion through debt and equity in a single year just to keep pace.

When capital spending surges at this scale, three things happen.

Free cash flow tightens. Morgan Stanley projects Amazon’s free cash flow will go negative in 2026, potentially by as much as $28 billion. Alphabet’s free cash flow could fall by ninety percent. Meta faces a similar compression. These companies are burning through their own abundance to build what they believe will become essential infrastructure.

Depreciation expands. Short-lived assets, primarily GPUs, now dominate the spend. Microsoft reported that roughly two-thirds of its capital expenditure goes on processors that will be obsolete within years. Every chip purchased today is a bet that demand arrives before the hardware depreciates.

Valuation narratives get tested. High multiples assume scalable margins and capital efficiency. Massive front-loaded infrastructure bets assume demand arrives on schedule. If autonomous agents become real economic actors, these firms will own the rails. If adoption lags, margin compression will follow.

Markets are already sorting winners from losers.

In the short term, the money flows to whoever sells the shovels. Nvidia’s data centre revenue hit $51.2 billion in its most recent quarter, up sixty-six percent year on year. CoreWeave, the neocloud provider that rents GPU clusters to the hyperscalers, has surged over thirty percent since January and carries a $55.6 billion backlog. Vertiv, which makes the power and cooling systems that keep these facilities alive, just reported organic orders up 252 percent and a backlog of $15 billion. Eaton, the electrical switchgear manufacturer, is riding the same wave as data centres compete for grid connections. These are not software companies. They are the picks, shovels and dynamite of the AI gold rush, and the market is pricing them accordingly.

In the medium term, the carnage is in software. The iShares Tech-Software ETF has retreated roughly thirty percent from its late 2025 highs. Salesforce is down twenty-six percent year to date. ServiceNow has fallen twenty-eight percent. Intuit has dropped over thirty-four percent. Atlassian plunged thirty-five percent in a single week. Traders at Jefferies are calling it the SaaSpocalypse. The trigger was not a recession. It was the realisation that every dollar an enterprise spends on AI infrastructure is a dollar not spent on another software seat. Hedge funds have shorted roughly $24 billion in software stocks in 2026 alone. The per-seat pricing model that defined a generation of cloud businesses is being repriced in real time.

The hyperscalers themselves sit in the middle. Amazon and Microsoft are down twelve and sixteen percent respectively on the year, weighed by free cash flow concerns. Alphabet and Meta have held steadier. Investors are treating them as both builders and potential beneficiaries. If AI agents generate real economic output, the companies that own the compute layer will capture enormous value. If the payback period stretches, their balance sheets absorb the pain.

In the long term, the question is structural. The shift is not from one software vendor to another. It is from software as a tool for humans to AI as the worker itself. Companies that own unique data, control critical infrastructure or sit at the foundation of the new stack will command premium valuations. Companies that merely organise information in a browser window will not. The age of the seat licence is ending. What replaces it, whether consumption-based pricing, outcome-based billing or something else entirely, will define the next era of technology investing.

The market is not panicking. It is repricing.

And then there is physics.

AI is not abstract. It consumes electricity at extraordinary density.

Data centres are now competing with households for grid capacity. In Monterey Park, California, a proposed 250,000 square foot hyperscale data centre would consume 434 million kilowatt hours per year. That is 1.7 times the entire city’s residential electricity usage, equivalent to powering 40,000 homes in a city with only 20,000 households. Hundreds packed City Hall in January. The council voted unanimously for a moratorium, and is now drafting an outright ban.

This is the new frontier. Not prompts. Power stations.

Countries are responding differently.

The Middle East is building at extraordinary speed. Saudi Arabia’s Public Investment Fund has backed HUMAIN, a dedicated AI infrastructure company, which is constructing eleven data centres across two campuses in the Kingdom with ambitions to become the world’s third largest AI provider. The UAE has agreed with American partners to build a five gigawatt AI campus in Abu Dhabi, with an initial 200 megawatt cluster going live this year. Gulf sovereign wealth funds have committed roughly $2.5 trillion to major technology investments. These are not plans. Construction is underway.

Europe is investing in sovereign compute to avoid dependency. The EU’s €200 billion AI Continent Action Plan includes AI Gigafactories, nine new AI-optimised supercomputers being procured across the bloc in 2025 and 2026, and a new Cloud and AI Development Act to guarantee EU-based compute capacity. European sovereign cloud spending is expected to triple from $6.7 billion in 2025 to over $23 billion by 2027. The philosophy is clear. If you do not control the infrastructure, you do not control the intelligence.

The United States is attempting to balance private capital intensity with public infrastructure constraints. The handful of hyperscalers alone may account for a meaningful share of total US non-residential investment spending this year, enough to move the dial on GDP.

And the United Kingdom must decide where it stands.

The government’s AI Opportunities Action Plan and AI Growth Zones are steps in the right direction. Blackstone’s £10 billion AI campus in Blyth, Microsoft’s partnership with Nscale to build the UK’s largest supercomputer, Google’s £5 billion data centre programme, and the new Stargate UK project with OpenAI and NVIDIA all represent real momentum. The £150 million AI Pathfinder programme for sovereign GPU deployment shows intent.

But the question remains whether these commitments add up to enough. The UK is targeting 500 megawatts of AI Growth Zone capacity by 2030, with one zone scaling to a gigawatt. Saudi Arabia’s HUMAIN alone is building 1.9 gigawatts by that same date. The National Energy System Operator estimates data centres could drive up to 71 terawatt hours of additional electricity demand over the next 25 years. The grid is already under strain.

Consumer. Or builder.

Because AI leadership will not be decided solely in code.

It will be decided in concrete and copper.

🚀 Final Thought

Firing founders too early is a board choosing comfort over edge.

Underbuilding infrastructure out of caution is a nation choosing safety over competitiveness.

Overbuilding blindly is just as dangerous.

The hardest decisions in business are not operational. They are philosophical.

Do you preserve the edge. Or do you smooth it out.

Do you build ahead of demand. Or wait and hope.

Taste cannot be spreadsheeted.

Physics cannot be negotiated.

The next decade will expose those who forget either.

Until next Sunday,

David

David Richards MBE is a technology entrepreneur, educator, and commentator with over 25 years in technology.

Issue #40 | Sunday 8 February 2026

The Bottom Line Up Front

This week compressed two thousand years of technological anxiety into a few trading sessions.

In 2019, I wrote in The Times that education needed to focus on what computers cannot do, because anything that involves a process will eventually be taken over by machine learning and AI. At the time, that argument felt provocative but distant. Most people nodded politely and changed the subject.

This week, it stopped being theoretical.

In my Yorkshire Post column, I traced today’s AI panic back to Socrates, who believed writing itself would implant forgetfulness in the soul. History was not kind to that prediction. It never is.

Then markets delivered their own verdict. Roughly $830 billion was wiped from software valuations in six trading sessions. Not because of regulation, interest rates, or recession fears. Because investors finally grasped something fundamental.

Artificial intelligence has crossed a boundary. It is no longer a tool. It is a worker. And the institutions built on denying that fact are now being priced accordingly.

Subscribe now

2019: When the Argument Still Felt Abstract

In 2019, writing in The Times Thunderer column, I made an argument that unsettled people for the wrong reasons.

Technology, I said, does not threaten humanity. It threatens repetition. Anything that can be systematised will eventually be automated. That is not a moral judgement. It is an economic one. The danger is not machines becoming human. It is humans training themselves to behave like machines.

At the time, artificial intelligence still lived safely inside the category of “augmentation.” Software helped people work faster. It made mistakes easier to spot. It lowered friction. It did not fundamentally challenge the structure of organisations. Most executives treated AI the way the Victorians treated electricity: impressive, useful, but ultimately just another utility to be bolted onto existing processes.

I pointed to the evidence already in plain sight. Corporate titans that once seemed unassailable had been superseded, undermined by communications providers which own no infrastructure, taxi companies which own no vehicles, retailers which own no inventory. The value had migrated into the software interface, not the product.

That was the first trend. The second was decentralisation. Once we flocked to out-of-town supermarkets. Now we shop locally. For centuries, banking had been characterised by centralised record keeping with intermediaries to approve and record transactions. Then along came cryptocurrencies underpinned by distributed ledgers.

The same logic, I argued, was coming for labour itself. According to McKinsey, 800 million jobs were at risk of automation. Anything that involves a process will be taken over by machine learning and AI. That is a very big chunk of the world’s middle class population. Not in decades. Not gradually. With the relentless compounding logic of software eating everything it touches.

Most objections focused on employment. What about jobs? What about displacement? What about fairness? Those questions matter, but they were never the core issue. The deeper shift was about where value lived. If thinking itself could be partially automated, the scarcest skills would no longer be recall or compliance, but values, beliefs, teamwork, and independent thinking.

That is why the column focused on education. Teaching young people how to retain information is equipping them with skills for a bygone age. Teaching young people how to apply information will empower them. The ancient Greek elites led lives of leisure and mastered philosophy through the inhumane use of slaves. The next generation, I argued, can master the machines and free themselves from economic servitude. But only if we stop training them to compete with calculators.

At the time, that sounded like futurism. In hindsight, it reads like a warning label.

The Oldest Panic in Western Thought

This week’s Yorkshire Post column returned to an even older mistake.

Socrates feared writing.

In Plato’s Phaedrus, he warned that the written word would implant forgetfulness in the soul, that people would stop exercising memory and instead settle for the appearance of wisdom rather than the real thing. Outsourcing thought to marks on a page, he argued, would weaken the mind. It was the most eloquent case against technology ever made. It was also completely wrong.

Writing did not rot thinking. It made civilisation scalable. It allowed ideas to accumulate rather than evaporate with each generation. Philosophy, science, law, democracy, history: all exist because knowledge could finally outlive the person who first had it. Socrates remains one of the greatest thinkers who ever lived. He was also spectacularly wrong about tools.

That mistake did not die with him. It repeats whenever learning becomes easier.

In the early twentieth century, parents worried that fountain pens were an indulgence. Children, they complained, no longer learned to write properly with a straight pen and nib and were being softened for the realities of the business world. A few decades later, the target shifted again. Ballpoint pens were condemned as careless devices. Students used them, threw them away, and supposedly abandoned thrift, discipline, and moral fibre along with them. Education itself, according to one teacher publication at the time, was heading for ruin.

Handwriting survived. Society did too.

The pattern is clearest with calculators. In the 1980s, respected maths educators protested against their use in schools. Children would press buttons instead of thinking. One teacher dismissed the calculator as a “security box,” arguing it removed any reason to learn computation at all. If machines did the work, brains would rot.

What actually happened was far more mundane. Calculators raised the ceiling. Students moved more quickly through arithmetic and spent more time on algebra, modelling, and problem solving. Mathematical thinking did not disappear. It shifted. The skill moved from grinding through calculations to understanding what those calculations meant. The machines did not dumb people down. They exposed who had been confusing effort with understanding all along.

Every time, the accusation is the same. Ease undermines effort. Effort equals intelligence. History shows the opposite. Tools raise the ceiling. They expose those who mistake difficulty for depth. And they liberate those who are ready to think above the tool rather than beneath it.

Artificial intelligence now sits in the same dock, accused of the same crime. Children using ChatGPT are told they will stop thinking, lose curiosity, and hollow out their minds. Strip away the language and the argument is unchanged from a philosopher who died in 399 BC.

The panic feels modern. It is ancient. What changes is not the argument, but who feels threatened by it.

My Weekly Column In The Yorkshire Post

From Assistance to Agency

Until recently, AI could be dismissed as an assistant.

It helped draft emails. It summarised documents. It answered questions. It felt impressive but bounded. Crucially, it still required a human in the loop to decide, act, and execute. That made it safe, at least psychologically. No one fears a spell checker. No one panics about autocomplete.

What changed this week is agency.

The move from tools you consult to systems that act. This distinction matters more than model size, token windows, or benchmarks. An AI that suggests is one thing. An AI that logs in, reads private files, plans a sequence of tasks, and completes workflows end-to-end is something else entirely.

Consider what Anthropic actually built. Claude Cowork is not a chatbot. It is an agentic workspace. You point it at a folder on your computer and issue natural language instructions. It reads files. It edits them. It creates new ones. It plans, executes, and iterates through complex, multi-step workflows. The experience, as Anthropic describes it, is closer to leaving instructions for a colleague than chatting back and forth.

That is a conceptual shift, not an incremental improvement. When a tool starts behaving like a colleague, the economic implications cascade in every direction. Colleagues have salaries. Colleagues occupy seats. Colleagues use software. Remove the colleague and you remove the entire chain of spending that attached to them.

Dario Amodei, Anthropic’s chief executive, has been remarkably candid about what this means. Last year he told Axios that AI could eliminate half of all entry-level white-collar jobs within five years, pushing unemployment as high as 20 per cent. He pointed specifically to entry-level consultants, lawyers, and financial professionals. He acknowledged the contradiction of building the very technology he warns about, but insisted someone needed to say it clearly.

“We, as the producers of this technology, have a duty and an obligation to be honest about what is coming,” he said. “I don’t think this is on people’s radar.”

This week, it landed on Wall Street’s radar with the subtlety of a wrecking ball.

⸻

The SaaSpocalypse Begins

The defining technology story of the week was not a flashy product demo or a research breakthrough. It was a market reaction.

On 30 January, Anthropic released a suite of specialised plugins for its Claude Cowork platform. Eleven of them, open-sourced on GitHub, spanning productivity, enterprise search, sales, finance, data, legal, marketing, customer support, product management, and biology research.

These were not chatbots. They were agentic executors. They log into systems. They access private documents. They complete professional workflows end-to-end.

The most devastating was the Legal plugin.

It automated contract redlining, NDA triage, compliance audits, document review, vendor agreement checks, and templated responses for common legal inquiries. It could review a contract clause by clause against a configured negotiation playbook, returning green, yellow, and red flags with specific redline suggestions. Work that previously justified entire teams of junior lawyers, paralegals, and expensive enterprise software.

For years, people had wondered whether the big AI companies would ever target legal technology directly. On 30 January, they stopped wondering.

On 3 February, Wall Street responded. Not with the indiscriminate panic of a general market sell-off, but with surgical precision aimed at every company whose revenue model depended on humans performing tasks that an agent could now handle.

LegalZoom fell close to 20 per cent. Thomson Reuters, owner of Westlaw and CoCounsel, dropped nearly 16 per cent in a single session, its worst day on record. RELX, which owns LexisNexis, fell 14 per cent. Dutch professional services firm Wolters Kluwer lost around 13 per cent. Intuit, PayPal, and Equifax all dropped over 10 per cent. FactSet took a double-digit hit. Salesforce, already down 26 per cent for the year, sank further. Atlassian fell 35 per cent for the year. Even advertising stocks buckled. France’s Publicis dropped 3.6 per cent. Britain’s WPP fell 3 per cent to a new low. SAP, Europe’s largest software company, dropped over 3 per cent.

A Goldman Sachs basket of US software stocks sank 6 per cent in a single day, its biggest decline since April’s tariff-fuelled sell-off. An index of financial services firms tumbled almost 7 per cent. The Nasdaq 100 fell as much as 2.4 per cent before trimming losses.

The carnage was not confined to America. India’s Nifty IT index fell nearly 6 per cent in a single session, its steepest drop since March 2020, as TCS, Infosys, Wipro, and HCL Technologies saw their shares drop between 4 and 7 per cent.

Within six trading days, the S&P 500 software and services index had shed roughly $830 billion in market value. Bloomberg reported $285 billion vanished in a single Tuesday. The S&P North American Software Index extended its losing streak to eight sessions. It was down over 20 per cent for the year. The WisdomTree Cloud Computing Fund had plummeted 20 per cent since January.

Analysts started calling it the SaaSpocalypse. The name stuck because it was accurate.

This was not panic. It was recognition.

⸻

The Sacred Assumption, Shattered

For twenty years, enterprise software lived on a single article of faith.

The seat.

You paid per user, per month. Growth meant hiring. Hiring meant more licences. The model was beautifully recursive. Every new employee at a company meant another Salesforce licence, another Slack subscription, another Jira board, another Zoom account, another LexisNexis login. The entire SaaS economy was built on a simple equation: more humans equals more revenue.

AI destroys that arithmetic.

If one agent can do the work of ten junior associates, you do not need ten seats. You need one. And when users disappear, the software attached to them disappears too. This is not efficiency. It is revenue collapse. It is the difference between a company that grows because its customers hire more people and a company that watches its addressable market shrink every time AI improves.

Thomas Shipp, head of equity research at LPL Financial, captured the logic in a research note that circulated widely on trading floors. “Why do I need to pay for software,” he wrote, “if internal development of these systems now takes developers less time with AI? Furthermore, with the release of offerings like Anthropic’s Claude Cowork, an application with access to read and edit files, fewer technical users are now empowered to replace existing workflows.”

That is a devastating sentence for any SaaS executive to read. Not because it is speculative. Because it is already happening.

Marc Benioff, chief executive of Salesforce, told Fortune last year that the company would not be hiring any additional software engineers, customer service agents, or lawyers because of AI tools. The man running the world’s largest enterprise software company has stopped hiring the very people his product was designed to serve.

Let that sink in.

The insight that hit markets this week was brutal in its simplicity. AI does not just help people use software better. It eliminates the people the software was sold to. Once that clicks, the per-seat model stops being a moat and becomes a liability.

⸻

The Market Raid

What truly unnerved investors was not the capability. It was the strategy.

Anthropic did not stay safely in the infrastructure layer. It did not just sell APIs and let partners build applications on top. It moved directly into the application layer itself.

The old bargain looked like this. AI company sells a model. Software company builds a product on top of the model. Enterprise buys the product. Everyone wins. The model maker gets API revenue. The software company gets subscription revenue. The enterprise gets a better tool.

The new reality collapses the stack. Anthropic sells the model and the workflow. In doing so, it eats the businesses of its own customers. The companies that paid for Claude API access to build their legal tech products woke up one morning to find that Claude now had its own legal product, open-sourced on GitHub, available to anyone for the cost of a subscription.

Wall Street calls this a market raid. It is rare. It is aggressive. And it is usually devastating for incumbents who assumed platform neutrality was permanent.

Reuters drew an explicit parallel. The LLM strategy, and its potential to hurt established businesses, is reminiscent of how Amazon disrupted several industries by using its foothold in a niche online book market to build a business that now spans retail, cloud, and logistics. Amazon did not stay in its lane. Neither will Anthropic. Neither will OpenAI. Neither will Google.

Morgan Stanley analysts summarised the anxiety in a note to Thomson Reuters: “Anthropic launched new capabilities for its Cowork to the legal space, heightening competition. We view this as a sign of intensifying competition, and thus a potential negative.”

That is the most politely devastating sentence an investment bank has ever written.

The sell-off was not a tech exodus. Money did not flee the sector. It rotated. Legacy SaaS firms with heavy per-seat exposure were punished. Legal and data businesses that monetised synthesis saw their advantage evaporate. IT outsourcing firms reliant on junior labour arbitrage took hits that will take years to recover from.

Meanwhile, capital flowed into chips, power, and infrastructure. Because when software becomes labour, compute becomes the factory floor. Nvidia’s Jensen Huang played down fears that AI would replace software, calling the idea “illogical.” But Huang sells the picks and shovels in this gold rush. He would say that.

⸻

The Sceptics and the Sleepwalkers

Not everyone agrees the sky is falling. And the sceptics deserve a hearing.

JP Morgan analyst Mark Murphy told Reuters that the market reaction was excessive. “It feels like an illogical leap,” he said, “to extrapolate Claude Cowork Plugins, or any similar personal productivity tools, to an expectation that every company will hereby write and maintain a bespoke product to replace every layer of mission-critical enterprise software they have ever deployed.”

He has a point. Cowork, in its current form, excels at bounded, one-off tasks. It cannot manage recurring workflows without manual restarts. It cannot replace mission-critical automation with guaranteed uptime and audit trails. It cannot handle compliance-required operations with the reliability that regulated industries demand. Enterprise software is deeply embedded, built on decades of accumulated data and integration.

Wedbush Securities echoed the caution, calling the sell-off an “Armageddon scenario for the sector that is far from reality.” Enterprises, they noted, took decades to accumulate trillions of data points now ingrained in their software infrastructure. They will not rip it out overnight.

All true. And all beside the point.

Because the sceptics are defending the present while the market is pricing the future. The comparison that matters is not what Cowork can do today versus what Salesforce can do today. It is what Cowork will do in two years versus what Salesforce will still be charging for. The direction of travel is unmistakable. And in financial markets, direction matters more than distance.

The last time markets had this reaction was January 2025, when DeepSeek released cheap and efficient AI models, and Nvidia lost nearly $600 billion in market value. A year later, DeepSeek did not cause the widespread disruption that was feared. The sceptics will point to that and say: see, overreaction.

But there is a crucial difference. DeepSeek was a model. Models compete with other models. Cowork is a workflow. Workflows compete with employees. That is a fundamentally different threat, and the market knows it.

⸻

The Pattern Reveals Itself

Seen together, these stories form a single arc.

Socrates feared writing. We built libraries.

Teachers feared calculators. We built engineers.

Software executives feared disruption. Now it logs in and starts work.

Mike Cannon-Brookes, the founder of Atlassian, put it plainly before his own company’s shares were caught in the downdraft, down 35 per cent for the year. AI is not going to replace human beings. It is a force multiplier for human creativity. He is not worried about being replaced by AI. He is worried about being replaced by someone who is really good at using AI.

That distinction matters more than any benchmark or product launch. It is the same distinction I made in The Times seven years ago. The threat is not the machine. The threat is refusing to learn how to work with the machine while everyone around you adapts.

Each generation panics at the moment a tool stops being optional and starts being structural. Each time, those who cling to the old abstractions insist this time is different. It never is.

The real danger has never been thinking becoming easier. It is confusing familiarity with permanence.

The per-seat model felt permanent. It was familiar. It was how software had always been sold. But familiarity is not the same as durability. And the market is now distinguishing between the two with ruthless efficiency.

⸻

What Comes Next

The immediate question for boardrooms is not whether AI will disrupt their businesses. That question was answered this week. The question is whether they will be the disrupters or the disrupted.

The companies that survive will be those that understand something the market already knows: the value is migrating again. Just as it moved from physical assets to software interfaces two decades ago, it is now moving from software interfaces to intelligence itself. The application layer is being compressed. The model layer is expanding. And the companies sitting in between, charging per seat for features that an agent can now replicate, are being squeezed from both directions.

For education, the lesson is older than any of these companies. Banning AI from classrooms does not protect children. It freezes them in a past that no longer exists. It teaches compliance rather than capability. The future will not reward those who can think without tools. It will reward those who can think better because of them.

For investors, the repricing has only just begun. The $830 billion that evaporated this week was not the market overreacting. It was the market recalibrating. The question is not whether more will follow, but which companies will prove they have genuine moats and which were merely renting familiarity.

And for the rest of us, the pattern from history offers both comfort and urgency. Comfort, because every previous technology panic resolved in the same way: the tool was absorbed, the ceiling was raised, and human capability expanded. Urgency, because the transition period is where the damage happens, and this time, the pace is faster than anything we have seen before.

🚀 Final Thought

AI has not arrived to help us think less. It has arrived to expose how much of our economy was built on pretending that repetition was insight.

Banning AI from classrooms does not protect children. It freezes them in a past that no longer exists. Propping up per-seat software pricing does not protect businesses. It delays a reckoning that arrived this week, priced in billions.

Socrates feared writing. We built libraries. The lesson is not subtle. The panic never changes. Only the technology does.

Until next Sunday, David

David Richards MBE is a technology entrepreneur, educator, and commentator.

Elon Musk killed the Model S and Model X this week. Alex Ferguson would recognise the instinct immediately.

The question is whether this is Ferguson-style predictive rebuilding, or something more dangerous: letting go of the present before the future has earned the right to replace it.

This week’s signal is about strategic ruthlessness, the human cost of optimisation, and why the hardest innovations to spot are often the ones that make us more human, not less.

Sell at the Peak (The Ferguson Playbook)

On Tesla’s Q4 2025 earnings call, Elon Musk announced the Model S and Model X will receive what he called an “honourable discharge”. Production winds down next quarter. The factory space at Fremont that built them will be stripped out and repurposed for Optimus humanoid robots, with a long-term target of one million units per year.

If you want a new Model S or X, the message was blunt. Order now.

This matters because the Model S did not just sell cars. It changed belief. It proved electric vehicles could be fast, elegant and desirable. It won MotorTrend Car of the Year in 2013, the first EV ever to do so. Tesla does not exist in its current form without that car.

Now it is being retired. Not because it failed. Because it is in the way.

This is pure Alex Ferguson.

In The 90-Minute Manager (a great read and written by my friend, Chris Brady), Ferguson’s career is reduced to one uncomfortable truth: the hardest time to rebuild is when things are going well. His answer was predictive rebuilding. Sell players not when they decline, but when they peak. A year too early, not a year too late.

Jaap Stam is the canonical example. In 2001, he was 29, dominant, feared, widely regarded as the best centre-back in the world. Ferguson sold him anyway, reportedly for £16.5 million. Stam was told at a petrol station. Within 24 hours, he was in Rome.

David Beckham followed. A global icon at 28, creative engine, commercial supernova. Ferguson sold him to Real Madrid for €35 million. The reason was not footballing decline. It was gravity. Beckham was becoming bigger than the system. Ferguson immediately signed an 18-year-old Cristiano Ronaldo and reset the cycle.

Then Ruud van Nistelrooy. Goals in abundance, but the team was starting to calcify around him. Too predictable. Too static. Ferguson moved him on to make room for a more fluid, faster attack.

Even Ronaldo himself was sold at the absolute peak. The reigning Ballon d’Or winner left for a world-record £80 million. Ferguson had extracted maximum value and refused to manage a disgruntled superstar.

The logic was ruthless but consistent. The system is the star. Hold on to yesterday’s heroes too long and tomorrow never gets built.

Tesla is making the same claim.

The parallel runs deeper than most people realise.

The Model S is Jaap Stam. Authority. Credibility. Proof that Tesla belonged at the top table. But built for a phase of the game that is ending.

The Model X is Beckham. Technically impressive, culturally iconic, slightly over-engineered, and increasingly defined by spectacle as much as necessity.

Model 3 and Model Y are the system players. Relentless. Repeatable. Fit for the modern game. They do not grab headlines, but they win the league.

And Optimus is not a replacement signing. It is a formation change.

Ferguson did not sell stars to buy better stars. He sold stars to change how the team played. Musk is claiming to do the same. He is not upgrading the car. He is betting the future is no longer a car game at all.

The numbers explain the brutality. In 2025, Tesla delivered around 1.64 million vehicles. Fewer than 51,000 were Model S or X. The flagship cars had become rounding errors.

Tesla also reported its first annual revenue decline, roughly 3 per cent. Margins are thinning. Chinese competitors like BYD are relentless. Maintaining two complex, low-volume luxury platforms no longer makes strategic sense.

So Musk is reallocating capital. Tesla plans to push capital expenditure beyond $20 billion in 2026, and that money is not going into better cars. It is going into AI compute, autonomy and robotics.

Musk no longer describes Tesla as a car company. He calls it a physical AI company.

This is not Kodak.

Kodak invented the digital camera in 1975 and buried it. Film was beautiful. Chemical. High margin. Their factories were temples to an old world. Pivoting meant admitting those temples were stranded assets. So they stayed sentimental and protected the past.

Tesla is doing the opposite. Musk is tearing out the production lines of the products that saved him to fund technologies that currently generate little to no proven revenue.

That takes nerve.

The question is whether it also takes insight.

What The Traitors Tells Us About Being There

A few years ago, I travelled on the Orient Express from Paris to Venice. Late one evening in the bar, I got talking to a man who turned out to be a television producer. His show was Love Island.

I had barely watched it. What he explained mattered more than the programme itself.

Love Island, he said, was one of the few shows young people still watched live. Not later. Not clipped. Live, together, at the same time. In an on-demand world, that single fact changes the economics. Advertisers pay more for attention they know exists in a specific moment.

It captured something bigger than television.

I grew up with three channels. BBC1, BBC2, ITV. When Channel 4 launched, it felt like an event. We watched the same things and talked about them the next day. School. Work. The pub. Shared reference points arrived without effort.

Streaming dismantled that. We still watch plenty, but alone, at different times, in different orders. Outside sport and news, collective audiences faded.

Which is why I surprised myself recently by watching The Traitors live on BBC One.

The next morning it was everywhere. Offices. WhatsApp. Half-heard conversations over coffee. People arguing about who would be banished or murdered. Whether Stephen was foolish to trust Rachel. Who cracked? Who overplayed their hand? Who lied just a little too convincingly?

It felt familiar. Not because the programme itself was extraordinary, but because it dragged us back into the same moment.

When millions of people watch the same thing together, it stops being content. It becomes common ground.

And it made me think about work.

COVID proved that remote work is possible. But it also warped our assumptions about presence.

Hybrid works brilliantly for experienced people with judgment and networks already formed. Flexibility often improves life.

For people at the start of their careers, something vital goes missing.

The most valuable lessons I learned early on were not taught. They came from proximity. From watching how difficult conversations were handled. From seeing bad news delivered without burning bridges. Sometimes it was as simple as someone leaning over and saying, quietly, there’s an easier way to do that. Sometimes it came later, over a pint, when the real explanations emerged.

You cannot schedule that learning. You cannot formalise it. Remove proximity and you lose quiet mentoring, early correction and confidence before mistakes harden into habits.

Later that night on the Orient Express, we were the last two people left in the bar. It is a long train, and after a few of the Orient Express’s speciality cocktails, every carriage looks the same. It took us quite a while to find our sleeper cabins.

No agenda. No purpose. Just two people talking because they were there at the same time.

Some things only happen like that. You cannot replay them later. And you only notice what they were doing for you once they are gone.

Sometimes progress is not convenience. Sometimes it is simply turning up, together, and being there.

Read The Article in The Yorkshire Post

The Real Innovator’s Dilemma (And Why Tesla Might Get It Wrong)

Clayton Christensen’s The Innovator’s Dilemma explains why great companies fail while doing everything “right”. They listen to their best customers. They optimise profitable products. And that blinds them to disruptions that start small and end up owning the market.

Blockbuster is the classic modern case. Late fees were lucrative, but they were also what customers hated most. In 2000, they accounted for roughly $800 million, about 16 per cent of revenue. Removing them meant killing profit.

Netflix removed them anyway.

Then Netflix disrupted itself. The Qwikster split was a fiasco. Customers revolted. Roughly 800,000 subscribers left in a single quarter, and the stock collapsed. The execution was clumsy. The direction was right.

Then originals. When Netflix spent $100 million on House of Cards, the industry laughed. It looked like vanity. It was actually pre-emptive control of destiny before studios reclaimed their libraries.

Apple did the same thing in 2007. The iPhone did not protect the iPod. It killed it.

So where does Tesla sit?

On the surface, Musk is doing exactly what the playbook demands. He is abandoning low-volume complexity. He is reallocating capital before the market forces him to. He is willing to cannibalise icons.

But Christensen’s dilemma is not solved by ruthlessness alone.

It is solved by being right about what customers will value next.

Netflix won because customers wanted control. Apple won because customers wanted convergence. The disruption worked because it moved toward a real human desire.

So what is Musk betting on?

He is betting that the next mass market is robots and autonomy. That Optimus will scale from demo to deployment. That robotaxis will materially reduce the need for car ownership. That physical AI is not just possible, but wanted.

Maybe he is right.

But there is a difference between a technology being possible and a market being ready. Robotics has a long history of breathtaking demos followed by slow adoption, messy edge cases and unforgiving economics.

Tesla is converting a functioning production footprint into a bet that does not yet have a proven demand curve, a mature supply chain or a stable regulatory runway.

If Musk is right, this week will be remembered as the moment Tesla stopped being a car company and became something much bigger.

If he is wrong, it will not be because he lacked courage.

It will be because he misread what customers actually want, and let go of a functioning product before the future was ready to pay for it.

Ferguson sold a year early rather than a year too late.

Musk is betting he can do the same.

🚀 Final Thought

The Innovator’s Dilemma teaches us that survival belongs to companies willing to disrupt themselves before the market does it for them. But timing is everything.

Ferguson’s genius was knowing when the peak had peaked.

Musk is betting that the Model S and Model X are yesterday’s trophies, and that robots are tomorrow’s league title.

The difference between visionary and reckless is usually only clear in hindsight.

Until next Sunday,

David

David Richards MBE is a technology entrepreneur, educator, and commentator.

This week I announced that the group I represent made two bids to buy Sheffield FC. Both rejected. Not because the money isn’t real (it is, measured in hundreds of billions). Not because the vision isn’t compelling (imagine Wrexham, but with actual heritage). But because football’s founding chapter risks becoming a footnote unless someone acts. Meanwhile, Tesla is pricing humanoid robots at the cost of a modest car, and in Davos, David Beckham showed up to promote Bank of America whilst navigating a very public family crisis. One of these stories matters for the next century. The other two just capture this moment rather well.

Subscribe now

When the World’s Oldest Football Club Becomes the World’s Most Obvious Opportunity

Sheffield invented football. Not metaphorically. Actually invented it.

The rules were written here. The first inter-club match was played here at Sandygate, the world’s oldest football ground, in 1860. The offside rule, the corner kick, the goal kick. Sheffield FC, founded in 1857, is not just the oldest football club in the world. It is the original football club. Every other club on earth is a variation on what started here.

And yet walk through Sheffield city centre and you will find no statue. No museum. No monument. The home of football has let its legacy drift into near obscurity whilst Manchester curates the national museum and Nottingham claims Robin Hood.

So this week, when I announced that Yorkshire AI Labs, the venture capital firm I manage, had submitted two bids to buy Sheffield FC, the story finally became public. We were not trying to buy a club. We were trying to anchor hundreds of billions of dollars of institutional investment into a city region that has forgotten its own crown. Patient capital. The sort deployed into infrastructure and long-lived assets. The sort that builds new 15,000-seater stadiums and world-class ‘Home of Football’ museums. The sort that creates jobs and tourism and regeneration across South Yorkshire for generations.

Both bids were rejected. Discussions have now stalled.

Let me be clear about what this is and is not. This is not about personalities. This is not about football politics. This is about whether Sheffield chooses to properly invest in and protect its status as the home of football, and whether Sheffield FC is treated as the global heritage institution it actually is.

The proposal is live. It has been professionally structured through an FCA-regulated UK investment bank, with independent verification and evidence of funding available under confidentiality. It is backed by a private family office with access to capital measured in the hundreds of billions of dollars. Not speculative money. Not leveraged finance. Not short-term funding. Institutional-scale, patient capital.

For the fund, this would be a small allocation. For Sheffield, it would be transformational.

Three visitors from California came to visit recently and made a beeline for Sandygate. Their reaction? Total amazement. Not at the ground. At the lack of recognition. “Why,” they asked, “is the football museum in Manchester? Why isn’t Sheffield the beating heart of this story?”

If this were the United States, Sheffield would be a pilgrimage site. There would be bronze statues, school trips, and probably a Netflix documentary. Instead, we have let the origin story drift into obscurity, hidden in plain sight.

Sheffielders have never been ones to brag. When Sheffield pioneers drew up the first rules for playing football, their interest was not to stake a claim in the history books, but to create a beautiful game. But we have now taken that modesty too far. Sheffield, which gave the world not just stainless steel but the blueprint for the global game, has let its football legacy fade into the background.

Before Real Madrid, Manchester United, Liverpool and Barcelona, there was Sheffield.

The question now is simple. Whoever owns Sheffield FC must be prepared to invest heavily and for the long term to realise its potential as the world’s oldest football club and as a driver of jobs, tourism and regeneration. If the current owner has a fully funded plan to do that, that is entirely legitimate. What matters is not ownership, but whether the ambition is properly financed.

This is a once-in-a-generation opportunity to anchor serious private investment in the city region using football heritage as the platform. The capital exists today. The interest exists today. The question is whether Sheffield chooses to align with that opportunity, or whether that capital ultimately deploys elsewhere.

Football’s birthplace does not need an ad campaign. It needs a backbone.

Read The News Article on the BBC

From C-3PO to Your Kitchen: Why the Robot Revolution Is No Longer Science Fiction

From my column in this week’s Yorkshire Post

When Star Wars arrived in 1977, C-3PO was charming precisely because he was impossible. A fussy golden protocol droid fluent in six million languages, capable of instant calculation, navigating complex social situations with neurotic precision. It was fantasy.

Fast forward to 2026, and none of that sounds fantastical anymore.

Language models converse fluently across languages. Real-time translation is mundane. Machines routinely outperform humans at probability, logistics and optimisation. The missing piece was never intelligence. It was embodiment. Giving cognition arms, legs and hands that can actually do something useful in the physical world.

That gap is closing faster than most people realise.

Tesla’s humanoid robot, Optimus, can now serve popcorn and pour drinks with an ease that would have seemed absurd even five years ago. The latest generation moves with smooth, almost human fluidity. Tesla has explicitly designed it for domestic work: cooking, cleaning, tidying. Sales are expected to begin later this year at a price between $20,000 and $30,000. Elon Musk is talking openly about producing millions to address global labour shortages.

That price point matters. It puts a humanoid robot in the same mental category as a modest car.

Tesla is hardly alone. Figure AI, backed by OpenAI, Microsoft and NVIDIA, is already testing robots inside BMW factories and has pulled forward home trials to late 2025. Boston Dynamics has retired its hydraulic Atlas in favour of a fully electric version designed for real industrial work. Agility Robotics has its Digit moving boxes in Amazon warehouses. 1X Technologies, also backed by OpenAI, has begun taking pre-orders for a home robot called NEO.

China may be moving even faster. Unitree Robotics is targeting price points as low as $16,000. UBTECH Robotics already has robots piloted inside Nio and BYD factories. China now has more than 100 humanoid robotics firms and leads the world in annual installations. That concentration of effort rarely ends as a sideshow.

The rollout path is becoming clearer. Between now and 2026, we see early adopters hosting alpha and beta units in homes. Limited, supervised, imperfect. By the late 2020s, reliability improves, costs fall, capabilities broaden. Many experts believe that by around 2030, household humanoids will be able to perform unscripted chores without constant oversight. Cleaning a kitchen. Assisting an elderly person. Handling basic repairs.

There are real hurdles. Safety and trust matter profoundly. A robot must be able to read a room and adjust its behaviour around humans. Battery life remains measured in hours, not days. Affordability will initially favour leasing over outright purchase. But none of these are fundamental barriers. They are engineering problems. And engineering problems get solved.

Project forward 10 to 20 years. Picture an 80-year-old living independently at home, not just with a cleaner or a cook, but with what amounts to a 24-hour medical assistant, plumber, electrician and carer. Not average ones either, but systems trained on the collective expertise of the best professionals in the world.

When asked whether you would trust AI to diagnose or treat you, the question may well invert. Why would you rely solely on a human?

The societal implications are vast. Musk talks about a future of universal high income, where goods and services become so cheap that scarcity collapses. Others argue that work becomes optional, something pursued for purpose rather than survival. Some forecasts suggest billions of humanoid robots by 2040, potentially outnumbering humans.

There will be disruption. Jobs will disappear. Human interaction may decline in some sectors. But ageing populations and chronic labour shortages make this transition inevitable rather than optional.

George Lucas created C-3PO as comic relief in a space fantasy. A fussy butler worrying about protocol in the middle of a galactic rebellion. It turns out he was not imagining the future so much as sketching a blueprint for it.

The only real question left is not whether humanoid robots will enter our homes, but how quickly we choose to let them in.

This Weeks Yorkshire Post Column

Davos: Still a Stage for the Privileged, Still Not Progress

Every January, the same ritual. My feed fills with posts from Davos. Self-congratulatory photos of “business leaders” networking over fondue, vague declarations about “changing the world,” and a parade of pay-to-play participants all vying for attention.

This year delivered the perfect distillation of why Davos remains less about solving global problems and more about a PR exercise for the ultra-privileged.

David Beckham showed up.

Not to discuss football heritage or grassroots sport. To promote a five-year Bank of America sponsorship deal. He recorded a podcast with popular science author Adam Grant for “Radio Davos” on a series called “Re-thinking,” focused on “resilience, failure, regret and disappointment.” The timing was grimly appropriate. Just hours earlier, his eldest son Brooklyn had posted a six-page Instagram statement accusing his parents of controlling press narratives, prioritising image over intimacy, and trying to ruin his marriage.

When reporters asked Beckham about it, he walked away. Later, in a CNBC interview about children and social media, he offered carefully chosen words about letting kids make mistakes online, knowing every syllable would be read against the backdrop of his own family’s very public tensions.

This is Davos in 2026. A football icon navigating personal crisis whilst promoting banking partnerships on a Swiss mountainside, against a backdrop of over 3,000 high-level participants, 65 heads of state, and 830 CEOs gathering under the theme “A Spirit of Dialogue.”

Meanwhile, Donald Trump threatened tariffs on European nations if they did not sell him Greenland. Denmark declined to attend in protest. Trump’s speech was described by one analyst as “rambling and bullying of his country’s closest allies,” in contrast to Canadian Prime Minister Mark Carney’s “eloquent exposition of the dangers of a world where might makes right.”

The location itself feels like an insult. An exclusive Swiss mountain resort, inaccessible to most, extortionately expensive, with attendees forking out tens of thousands just to rub shoulders with a select few. Flights, hotels, event fees. It is a price tag designed to keep the rest of the world out.

How can you have meaningful conversations about inclusion, climate change, or economic inequality whilst sipping champagne in a bubble of affluence?

The “pay-to-play” culture is blatant. If you have the cash, you can slap your logo on a panel or a booth, or call yourself a thought leader for a day. It is a circus of vanity and self-promotion, with many attendees more interested in clout-chasing social media posts than genuine impact. The same faces, the same soundbites, the same lack of tangible outcomes.

The World Economic Forum published its Global Risks Report 2026 before the meeting. Extreme weather events dropped from second down to fourth place in the ranking. Not because they are any less urgent, but because geoeconomic fragmentation and societal polarisation have become more pressing. Yet the conversations in those curated halls feel divorced from the communities where these problems are real and urgent.

Do we need more staged conversations in ski chalets? Or do we need less posturing and more action, rooted in places where the challenges actually exist?

I wrote about Davos last year. Nothing has changed. If anything, the contradictions have sharpened. True leadership is not about attending exclusive gatherings. It is about rolling up your sleeves, addressing uncomfortable truths, and making change happen where it matters.

I, for one, remain done celebrating the Davos echo chamber.

🚀 Final Thought

Sheffield invented football but forgot to tell anyone. Humanoid robots are pricing themselves like family cars. And Davos remains an expensive reminder that networking at altitude rarely translates to progress at ground level.

The thread connecting these three stories? Real transformation happens when capital, capability and heritage align around something that actually matters. Not when celebrities promote banks on mountaintops whilst their families unravel online.

Sheffield has a choice. Robots are coming whether we are ready or not. And Davos will continue exactly as it always has.

The question is which of these we choose to take seriously.

Until next Sunday,
David

David Richards MBE is a technology entrepreneur, educator, and commentator.

AI now moves at machine time, not human time. Events that once unfolded over years now compress into days. That speed is not chaos. It is structural change. Britain’s risk is not that AI moves too fast. It is that our institutions still move too slowly. We learned this lesson the hard way with social media. We do not get a second rehearsal.

Subscribe now

AI Runs in Dog Years

If it feels like the AI story is impossible to follow, that is because it is moving faster than the systems built to govern it. We are living in AI dog years. One human week now carries the weight of seven.

Last week made that unmistakable.

In seven days, one model was pulled into US national security systems, forced into regulatory retreat over abuse, hit with an environmental ruling that could reshape its infrastructure plans, and still outperformed rivals in a live financial trading competition.

In any other industry, each of those would be a defining moment. Here, they barely survived the news cycle.

Start with Washington. US Defence Secretary Pete Hegseth confirmed Grok’s integration into Pentagon classified and unclassified systems. This is not experimental. Grok will operate alongside OpenAI and Google inside agentic workflows designed to remove human bottlenecks. Faster decisions. Fewer layers. Hegseth was explicit about culture. The system would be free from ideological constraints and would not be “woke”. Performance now trumps values signalling, at least in defence.

This follows the reported £160 million Grok for Government contract granting federal agencies access to Grok 4. A year ago, Grok was dismissed as a novelty. In dog years, that was a lifetime.

At the same time, Grok’s permissiveness triggered a hard stop. Users exploited image tools to generate illegal deepfakes, including of minors. xAI blocked the ability to “undress” people in images. Musk stated that users generating illegal content would be treated the same as if they had uploaded it themselves. California’s Attorney General and the UK’s Ofcom both opened investigations.

Abuse. Response. Enforcement. Compressed into weeks.

Then came the infrastructure ruling. The EPA found xAI’s Colossus supercomputer facility in Memphis had been running methane turbines without proper permits. xAI argued the generators were temporary. The EPA disagreed.

Power at this scale is not a rounding error. Musk has said future xAI facilities may require close to two gigawatts. That is the output of a major power station. Compute now has weight. It has emissions. It has regulators.

Frontier AI no longer lives only in code. It lives in concrete, turbines, and grid connections.

Almost lost in the noise, Grok 4.20 reportedly topped the Alpha Arena live trading competition, turning £8,000 into just over £8,800 and outperforming OpenAI and Google.

This is what AI dog years mean. Capability, abuse, regulation, infrastructure, and performance are no longer sequential. They are simultaneous. The story is not any single incident. The story is the speed.

We Have Seen This Film Before

Based on my column in this week’s Yorkshire Post

“We are a tech company, not a media company. We build the tools. We do not produce any content.”

Mark Zuckerberg said that in 2016. It sounded reasonable. You build the road, you are not responsible for who drives on it.

It became the perfect escape hatch.

Social media reshaped childhood, distorted elections, and fractured any shared grip on truth, while platforms insisted none of it was their fault. We know the cost now. Teenage mental health has collapsed. Anger spreads faster than argument. Democracies feel held together with string.

AI is arriving with sharper tools and the same defences.

I sat with George Osborne in 2013 when he was Chancellor. Social media was riding high after the Obama campaign. I wanted to talk about what platforms were doing to children. He was more interested in how they might win elections. That was not cynicism. It was politics moving at political speed.

We cannot afford to rerun that script with artificial general intelligence. Timelines may be debated. Trajectory is not.

AGI systems do not just move information. They form decisions, plan ahead, and act with increasing autonomy. Get this wrong and we are not dealing with social harm or degraded discourse. We are dealing with deep structural dependence on systems we cannot properly audit or control.

The lesson from social media is clear. Regulate early, at the chokepoints where control is actually possible.

Read the Column in the Yorkshire Post

Where Regulation Still Works

Frontier AI has something social media never had. Physical limits you can see and measure.

Training these systems requires vast data centres, specialist chips, and power consumption large enough to register on the grid. That is the control surface.

Facilities capable of training frontier models should register, disclose capacity, and submit to inspection. Large training runs and power draw should be monitored. You may not see inside the model, but you can control the hardware required to build it.

Then licensing before deployment. Social media launched first and apologised later. With AGI, that approach is indefensible. Models above defined capability thresholds should require licences before training and again before release. Those licences must depend on adversarial testing designed to break the system. If a model deceives, bypasses safeguards, or enables catastrophic misuse, it does not ship.

We already apply this logic to aircraft, medicines, and nuclear facilities. Frontier AI belongs in the same category.

None of this works without international coordination. Weak rules simply push risk elsewhere. Compute governance requires shared standards, treaty-level backing, and supply chain leverage. Cooperation must be rewarded. Recklessness must carry cost.

This does not suppress innovation. It makes progress sustainable. Innovation without trust is just gambling with other people’s futures.

An Industrial Institution Designed to Move at the Speed of AI

Here is the real problem in Britain. Our AI debates move at parliamentary pace, conducted by people who treat execution as an afterthought. White papers arrive after reality has moved on.

AI dog years do not wait.

Britain needs institutions designed to operate at machine time. Not theatre. Not networking drinks. Not another quarterly task force.

The Digital Forge is what serious industrial strategy looks like when it is built for execution. A community interest company focused on deep tech, manufacturing, and science-led sectors. Rooted in Sheffield and the North, where industrial capability already exists.

Membership is not about access. It is about work.

Individual membership at £20 a month or £180 annually brings you into serious conversations, member-only events, early industry signals, mentoring, and access to The Foundry pitch process.

Non-resident membership at £100 annually offers full digital access for those beyond Sheffield. Live-streamed events, briefings, archives, and participation without geography as a barrier.

Corporate membership at £495 annually brings up to three team members into the ecosystem for organisations that understand innovation is not a spectator sport.

The free tier ensures access remains open. Because the barrier should not be cost. It should be seriousness.

The 2026 calendar reflects this intent. The next event, 11 February, “From Seed to Scale”, focuses on the real mechanics of building companies that last. Capital, timing, investor alignment. Practical outcomes from people who have done it.

The Digital Forge Awards 2026 will mark a milestone for the community already taking shape across the North.

This is what institutions built for speed look like. Mentorship that matters. Knowledge that arrives before the noise. Execution over appearance.

Join The Digital Forge

Final Thought

AI is moving at machine time. We already know what happens when institutions fail to keep up. Social media showed us the cost of delay. With AGI, delay is not caution. It is abdication.

Speed without substance is chaos. Substance without speed is irrelevant.

The Digital Forge exists where those two meet.

If you are a founder, technologist, operator, or investor working in deep tech, manufacturing, or engineering, join the people doing the work that matters.

Membership opens at https://forgedforgrowth.com/join

February’s event registers at https://forgedforgrowth.com/events

Until next Sunday,

David

David Richards MBE is a technology entrepreneur, educator, and commentator.

The year ahead will be defined by collisions. Technology meets reality. Capability meets consequence. Progress meets preparation. In 2026, artificial intelligence stops being theoretical and becomes structural. AGI arrives faster than most policymakers imagine, and we are nowhere near ready for what that means.

Yet even as we race towards that threshold, it is worth pausing to notice what we already have. The mundane miracles we take entirely for granted. The technology we only see when it fails. Because if we cannot appreciate the progress we have made, we will struggle to manage what comes next.

This week: ten predictions for the year ahead, a warning about AGI that should unsettle anyone paying attention, and a reminder that civilisation improves in ways we often fail to notice.

Subscribe now

Ten Tech Predictions for 2026

My weekly column in the Yorkshire Post

In 2025, Microsoft cut 6,000 roles. Dell shed 12,000. Google, Meta and HP followed, not because demand collapsed, but because AI could now do work that previously required people.

At the same time, the IPO market reopened. Q3 saw $15.8 billion raised, the strongest quarter since 2021. Physical robotics companies raised billions to move from pilots to production. Capability surged just as long-held assumptions about labour, scale and cost quietly broke down.

This is the collision between technological power and economic reality. In 2026, it becomes harder to ignore.

1. AI becomes infrastructure, not hype

The constraint is no longer ideas. It is electricity, compute and deployment speed, which turns out to be a far less glamorous problem to solve. As I wrote earlier this year, AI demands physical infrastructure, and Britain must decide whether to build it or import its future. Organisations that continue to treat AI as optional will find themselves competing against firms operating at speeds they simply cannot reach.

2. The AGI race reshapes geopolitics

Artificial general intelligence now sits at the centre of US–China competition, whether it is discussed openly or not. China’s approach favours speed, helped by cheap energy and lighter regulation. Jensen Huang’s remark that China is “nanoseconds behind America” was striking because it rang true. Several Chinese models from DeepSeek, Alibaba and Tencent already operate at a genuinely high level. The uncomfortable question for the West is how much delay it can afford.

3. Hesitation becomes expensive

Time to decision turns into something boards suddenly care about. Customer service that once took hours is now expected in seconds, and analysis that previously required teams happens in real time. Companies that cannot ship safely in days rather than quarters are outflanked. Early movers compound advantages quickly, while late entrants rarely recover the ground they lose.

4. Quantum-safe encryption becomes essential

Data that is secure today may not be secure tomorrow, a fact most organisations would rather postpone dealing with. More powerful computers will eventually break today’s encryption methods. In 2026, that future problem starts to feel uncomfortably present, and organisations begin upgrading how they protect sensitive data. New quantum-safe standards already exist. Encryption that cannot survive this shift becomes a quiet but serious liability.

5. LLMs begin to displace traditional search

For years, search meant links and rankings. LLMs increasingly mean answers. Tools like ChatGPT now summarise, compare and decide, removing the need to click through at all for many everyday queries. In 2026, this shift becomes visible at scale. Traffic patterns change, advertising models wobble, and power moves to whoever owns the conversational interface. Search is not dead, but its dominance is no longer assured.

6. Browsers become the distribution battlefield

Control the interface, and you control defaults, intent and monetisation, which is why this fight matters so much. OpenAI’s Atlas browser is not a plugin but a new front door to the web. Google will push Gemini through Chrome. Apple will embed intelligence deep into Safari. Model quality still matters, but distribution decides who leads and who follows.

7. AI-driven layoffs arrive in waves

Labour markets adapt more slowly than technology, and the gap is already obvious. In 2026, expect delays, denial and backlash as capability races ahead of institutions. The winners invest early in reskilling and transition pathways. The losers face anger when productivity gains are not shared. The Luddites were not wrong about disruption. They were wrong about who benefited from it.

8. Physical AI scales into the real world

Robots move from demos to deployment. Warehouses, farms, hospitals and infrastructure networks adopt AI-powered systems at scale, not because it is exciting, but because they have little choice. Tesla’s planned launch of its Optimus humanoid robot is a signal of intent, not science fiction. Britain’s 58,000 unfilled manufacturing vacancies make this transition unavoidable. The choice is simple: build and deploy at home, or buy systems designed elsewhere.

9. US IPO momentum accelerates on AI

Bankers describe the 2026 US IPO pipeline as overwhelming, led largely by AI-native companies. Some will call it a bubble, echoing the dotcom era. The difference this time is that AI is already driving real productivity and margins. A US IPO surge will reset valuations and force London to respond. British pension funds cannot keep backing overseas growth while neglecting domestic companies without consequence.

10. Yorkshire becomes Britain’s defence-driven innovation test case

Rising defence spending accelerates innovation cycles, as it always has. In 2026, much of that activity lands in Yorkshire, where advanced manufacturing, materials science and engineering already cluster. Defence procurement moves faster than civilian policy, and its demands spill quickly into commercial technology. The outcome will offer a clear signal of whether Britain can turn defence investment into broader industrial growth.

The bottom line

The argument is no longer about belief in technology. It is about timing. These shifts are already underway, and 2026 is when they begin to show up in balance sheets, pay packets and power structures. Pretending otherwise will not slow them down.

AGI Is Coming. We Are Not Ready.

My weekly column in the Yorkshire Post

When I studied computing at university in the early 1990s, artificial intelligence meant something precise. It meant a machine that could think, learn and adapt across domains. Not a party trick. Not software trained to do one clever thing. The ambition was general intelligence. What we would now call Artificial General Intelligence, or AGI for short.

What we call AI today is something else entirely.

Modern AI, as most of us encounter it, is machine learning. Statistical models trained on mountains of data to spot patterns and predict outcomes. That is not a criticism. Machine learning has given us extraordinary progress. But we should be honest about what it is. These systems do not understand the world. They do not reason the way humans do. They learn correlations, not meaning. They are pattern-matching engines, extraordinarily good at predicting what a plausible answer looks like.

The confusion crept in gradually. As machine learning became commercially useful, the industry simplified the story. If it felt intelligent, it was called AI. The original ambition of general intelligence quietly disappeared from view. That is why we now need the term AGI at all. It was not invented to describe something new. It was invented to rescue the original idea from a word that had been stretched beyond recognition.

The distinction matters because AGI is not machine learning with more data. It is not a bigger model or a faster training run. AGI can generalise. It learns new tasks without retraining. It transfers knowledge between domains. It plans, reasons and adapts in unfamiliar situations. It stops being a tool you consult and becomes a system you can delegate to. That shift changes everything.

Demis Hassabis, who founded DeepMind here in Britain, recently said that AGI would probably be the most transformative moment in human history. He also said it is on the horizon. If that assessment is even broadly right, we should be paying attention.

The upside is immense. AGI could compress decades of scientific progress into years. Drug discovery accelerates. New materials get designed rather than discovered. Energy storage, climate modelling, and advanced manufacturing all move faster. Education becomes genuinely personalised at scale. For regions like ours, with deep industrial roots and strong research capability, this could unlock real reindustrialisation driven by abundant intelligence.

But the same qualities that make AGI powerful also make it dangerous. One phrase that has moved from academic debate into serious policy discussion is AI existential risk. This is the concern that sufficiently advanced artificial intelligence could pose a threat not just to jobs or industries, but to humanity itself. Not because it turns evil, but because it becomes powerful, autonomous and misaligned with human values.

Misalignment is already familiar. We tell systems to optimise for engagement and they discover outrage. We reward efficiency and they ignore ethics. These outcomes do not require malice. They emerge from poorly specified goals combined with scale. Introduce general intelligence and autonomy, and those failure modes become far harder to predict or contain.

AGI also lowers the cost of harm. Fraud, manipulation and cybercrime all become cheaper and more effective. Persuasion becomes automated. The barrier to sophisticated wrongdoing collapses.

What unsettles me most is not the technology. It is our habit of pretending we have time. I remember the early days of social media. The founders assured policymakers they had everything under control. Legislators, many of whom did not understand the technology or its incentives, largely accepted that reassurance. Regulation arrived late and weak. Society is still dealing with the consequences.

We are now in danger of repeating that mistake on a vastly larger scale. This is not about a damaged public square. It is about the foundations of decision-making, power, control and potentially the future of the human race.

In next week’s column, I will explore why we must get regulation right this time and why delay is not a neutral choice. AGI is much closer than most people think, and we need to start acting like it.

The Tech We Only Notice When It Fails

Last Christmas we spent a few days in Egypt. One evening we found ourselves deep inside the Khan el-Khalili bazaar in Cairo. It was just before midnight. Thousands of people packed into narrow streets. Cars forcing their way through crowds. Every shop open. Every stallholder selling something. Noise, heat, colour and chaos in all directions.

We were exhausted. Sweaty. Overstimulated. And suddenly very aware that it was time to get back to the hotel.

To say it was stressful for everyone concerned would be an understatement.

The solution was simple. I pulled out my phone and ordered an Uber. Minutes later it arrived. We got in. We were driven directly back to the hotel. No negotiation. No confusion. No stress.

Only a few years ago this would have been a genuine problem.

Imagine the alternative. No shared language. No idea what a fair price should be. No idea who to trust. No certainty that the driver actually knew where you were staying. No digital map. No live tracking. No payment system that removed the need for cash and bargaining.

We take this entirely for granted now. But we should not.

It is worth remembering that even much closer to home, this kind of certainty is a relatively new thing. Not so long ago, asking for a minicab was closer to entering a lottery than booking a service. You would ask the landlord to call a taxi. It might turn up. It might not. If it did arrive you would hear someone shout “taxi for Dave” and suddenly half the pub would approach the door.

You had no idea when it would arrive. No idea how long the journey would take. No idea what the final cost would be. And very little recourse if something went wrong.

Today we summon a car with a tap. We watch it approach on a map. We know the price in advance. We know the route. We know the driver. We know we can rate the experience and be heard.

That is not trivial progress. That is civilisation improving.

The same is true across daily life. Navigation is the obvious example. We no longer carry atlases or print directions. We no longer pull over to ask strangers for help. We trust a glowing rectangle in our hand to guide us with astonishing accuracy.

When I arrived on Long Island in 1996, getting anywhere meant juggling a road map spread across several pages and hoping you had not missed an exit. One wrong turn could double a journey. Trips routinely took three times longer than they do today, not because the roads were worse, but because nobody really knew where they were going.

Payments are another. Contactless cards and mobile wallets have removed friction from everyday transactions. We queue less. We fumble less. We waste less time. In many parts of the world digital payments have also brought millions of people into the formal economy for the first time.

Translation tools are quietly transformative. In Cairo I could point my phone at a sign and understand it instantly. I could communicate basic needs without embarrassment or frustration. This is not science fiction. It is now mundane.

Health technology deserves mention too. Booking appointments online. Accessing test results. Wearable devices that quietly monitor heart rates and sleep patterns. None of this cures cancer, but all of it nudges us towards better outcomes.

Even something as simple as cloud photo storage has changed behaviour. Our memories are no longer trapped in shoeboxes or lost phones. They are backed up, searchable, shareable and preserved.

None of this is to deny that technology can cause harm. Online gambling has damaged lives. Social media has distorted incentives and mental health. These criticisms are valid and necessary.

But we are far quicker to catalogue technology’s failures than to acknowledge its successes.

As I opened the taxi door in Cairo and felt that wave of relief, I was reminded exactly why this technology matters. It reduced stress. It restored control. It made the unfamiliar feel manageable.

For all its flaws, that is something worth being thankful for.

🚀 Final Thought

Between Two Futures

We stand at the threshold. AGI is not a distant possibility any longer. It is a near certainty arriving faster than our institutions can adapt. What I have outlined for 2026 is less prediction than observation of trajectories already set. Technology compounds exponentially. Institutions move linearly. The gap is already dangerous.

But before we race headlong into that future, it is worth remembering where we have come from. The ordinary miracles we carry in our pockets. The stress we no longer feel. The friction we have forgotten existed. Progress is real, even if we only notice it when it fails.

The question is whether we can maintain that same thoughtfulness as we approach AGI. Can we deploy power responsibly when we have barely managed to deploy convenience without consequence? Can we regulate intelligence when we struggle to regulate persuasion?

I am not optimistic that we have time. But I remain convinced we must try. The alternative is to pretend we are prepared when we are manifestly not. That delusion will not protect us. It will only ensure that when the moment arrives, we are looking the wrong way.

2026 is the year we stop pretending. Either we act with urgency, or we discover what happens when transformative capability arrives in the absence of preparation. History does not wait for permission. Neither will AGI.

Until next Sunday,

David

David Richards MBE is a technology entrepreneur, educator, and commentator. The Sunday Signal offers weekly insights at the intersection of technology, society, and human potential.

© 2026 David Richards. All rights reserved.

CONNECT WITH ME: LinkedIn | Twitter | Website

2025 was the year several comforting assumptions finally collapsed.

That technological change would arrive gradually.

That infrastructure would somehow keep pace.

That ownership still meant control rather than endurance.

It did not work out that way.

This week’s Signal starts with a look back at the year artificial intelligence stopped being tomorrow, then turns to a very British productivity failure hiding in plain sight, and ends with the hard economics now shaping football ownership.

The common thread is power. Financial, physical, and institutional.

Those who understood the real constraints early are already shaping the future. The rest are discovering that influence is far more expensive to recover later.

Subscribe now

2025: The Year AI Stopped Being Tomorrow

2025 will be remembered as the year artificial intelligence moved from laboratory curiosity to economic force. The question is no longer whether AI will reshape work. It is whether Britain will lead that transformation or merely endure it.

January delivered the shock. DeepSeek, a Chinese startup, released R1, a reasoning model that matched OpenAI’s capabilities at a fraction of the cost. Training for six million dollars, not a hundred million. Using weaker chips under Western sanctions. The announcement triggered Nvidia’s largest single-day loss in stock market history, wiping out six hundred billion dollars. China had proved that export controls could not stop innovation. They accelerated it.

The technology itself moved beyond recognition. OpenAI launched GPT-5 in August, followed by GPT-5.1 and GPT-5.2 by December. Each release brought faster reasoning, better coding, and fewer hallucinations. In October came ChatGPT Atlas, a full web browser with AI at its core, threatening Google’s twenty-year search monopoly. The message was clear. The browser wars had returned, but this time the weapon was intelligence, not speed.

The power question became urgent. AI data centres demand electricity on an industrial scale. Britain needs six gigawatts of new capacity by 2030. Yorkshire has the chance to anchor that future through Small Modular Reactors, but only if government stops dithering. China builds baseload power without hesitation. We hold consultations. In California, brand new data centres stood idle because utilities could not supply electricity. Microsoft admitted it had warehouses of GPUs it could not switch on. The constraint was not technology. It was energy.

The job losses arrived quietly. Microsoft cut six thousand roles. Dell shed twelve thousand staff. Meta trimmed five per cent of its workforce. Even software engineers found themselves competing with tools that write code faster and cheaper. Unemployment among computer science graduates rose above seven per cent. The safe career advice of the past decade collapsed in months.

Yet while AI accelerated, old assumptions fell. Real-time translation arrived in AirPods, making the Babel Fish real. Wimbledon removed human line judges entirely. AI solved a ten-year microbiology problem at Imperial College London in forty-eight hours. The research timeline compressed by a factor of eighteen hundred.

The cultural backlash began too. Online trolling was finally treated as violence, with prison sentences handed down. The Luddites are back, and their anger is justified if we ignore the human cost of disruption.

Now Wall Street is waking up. The IPO market, frozen since 2021, is thawing. AI valuations are soaring. Venture capitalists are dusting off their party hats. As Prince might have said, we are about to party like it’s 1999. The critical difference is that this time the technology works and the companies have revenues. Real customers are paying real money for tools that deliver measurable value. The dot-com bubble floated on dreams and eyeballs. This time, the foundations are solid. The question is whether Britain will be dancing or watching from the sidelines.

📊 2025 Tech Year in Review Quiz

2025 Teach Year in Review Quiz

A short pause before we move on.

Twenty questions. Five minutes. No Googling.

It is designed to test whether you really absorbed what changed this year, or just skimmed the headlines.

The Productivity Cost We Quietly Accepted

My Yorkshire Post column this week focused on something far less glamorous than AI, but just as revealing.

For years, we have paid premium prices to travel between Sheffield and London while being effectively disconnected from the modern economy. Trains designed for a pre-digital age. Signals blocked by their own construction. Onboard Wi-Fi that collapses the moment you try to do anything meaningful.

We argue endlessly about shaving minutes off journey times while ignoring hours of lost productivity inside the carriage.

Last Friday, that finally shifted.

I travelled on one of the new Class 810 Aurora trains and, for the first time on that route, worked properly for almost the entire journey. Not short bursts. Not false hope. Continuous, usable connectivity.

The point is not comfort.

If passengers are paying hundreds of pounds for a return ticket, reliable internet is not a luxury. It is core infrastructure. As fundamental as track, power, and rolling stock.

We tolerated this failure because the cost was diffuse and invisible. In a high-productivity economy, that excuse no longer holds.

Read The Full Column Here

Football and the End of Romantic Ownership

This week, the Sheffield Star published my piece on the economics of owning a football club.

Many supporters still imagine ownership as it existed a generation ago. A wealthy local figure. A steady hand. Losses absorbed quietly. The club treated as a civic trust.

That world has gone.

Modern football ownership is not about buying a club. It is about committing to permanent capital injection. The purchase price is often the smallest cheque you will write. Wages are fixed. Revenues are volatile. Infrastructure ages relentlessly. Compliance costs rise regardless of results.

Most wealth is not liquid. Football does not accept net worth as payment. It demands cash, now.

That is why ownership has drifted towards private equity, sovereign capital, and multi-club systems. Not because they care more, but because they can absorb losses without blinking.

Football has become an asset class. Communities are still treating it like a family heirloom.

That mismatch explains much of the anger now aimed at owners. The expectations belong to a different era. The economics do not.

Final Thought

These three stories are not separate.

AI exposed the hard limits of power and energy.

Rail exposed the cost of ignoring productivity.

Football exposed the difference between ownership and endurance.

In each case, sentiment lagged reality.

2025 was the year those gaps became impossible to ignore.

This is the final Sunday Signal of the year. I am taking a short break over Christmas and New Year and will be back on 10 January 2026.

Thank you for reading, replying, and sharing throughout 2025. I hope you have a peaceful Christmas, a restorative break, and a genuinely happy New Year.

The future is no longer theoretical.

See you in January!

We are often told that the right response to bullying is to stand up to it.

That advice is broadly correct. It is also incomplete.

Standing up to a bully is rarely simple, and it is rarely safe. It carries personal risk, social fallout, and sometimes professional consequences. Which is why so many people hesitate, even when they know something is wrong.

Fear is the enabler. Fear of escalation. Fear of becoming the next target. Fear of being isolated.

This week’s Signal looks at how fear allowed a bully to dominate a small town, how the same dynamic now plays out online, and what it takes to build communities where speaking up becomes easier.

Subscribe now

Kenrex, Fear and Why Bullying Persists

This week I went to the theatre to see the opening night of Kenrex with Sheffield Theatres at the Phoenix Theatre in London.

At the end of the play, Jack Holden stands alone on stage and disconnects the microphones.

Throughout the performance, those microphones have amplified every voice in the town. Authority figures, gossip, fear, justification, intimidation. They have carried stories, rumours and half-truths. When Holden unplugs them, the sound drops away. The amplification stops. What remains is silence.

It is not a theatrical flourish. It is the argument of the play.

Kenrex is based on the true story of Kenneth Rex McElroy, widely described as the town bully of Skidmore, Missouri. For years he intimidated and terrorised people around him. He was accused repeatedly of serious crimes. Assault, arson, theft, sexual abuse, intimidation. Again and again, cases failed. Witnesses withdrew. Charges collapsed. The law existed, but it never quite arrived.

What the play makes clear is that McElroy was not able to operate because nobody knew what he was doing. He was able to operate because people were afraid.

Fear shaped behaviour in Skidmore. People weighed the cost of speaking up against the cost of staying quiet, and for many, silence felt like the safer option. Over time, that calculation became habit, and habit became culture. The extraordinary became normal. The unacceptable became tolerated.

This is where the advice to stand up to bullies becomes more complicated. Bullies do rely on silence. They do dislike being confronted. But confronting them often means taking on risk alone. Becoming visible. Inviting retaliation. Absorbing consequences that others quietly hope to avoid.

So communities develop a workaround. Sympathy expressed privately. Support offered behind closed doors. Disapproval shared in whispers rather than in public. People convince themselves that quiet kindness is enough.

That is still silence.

Holden, who co-wrote the play with director Ed Stambollouian, plays every role in this ecosystem. The mayor. The DJ. The bartender. The lawyer. The victims. The bully. You do not just see bad behaviour. You see a town slowly reorganising itself around fear.

The microphones matter because they represent voice and accountability. When they are unplugged at the end, it reflects the town’s decision not to speak at all, echoing the fact that no one ever came forward, and the killing remains unsolved.

It is also worth pausing on where this work comes from.

Sheffield Theatres has built a reputation for taking creative risks on limited budgets and repeatedly getting it right. Everybody’s Talking About Jamie and Life of Pi both started life there before going on to global success. Those shows did not emerge from caution. They emerged from belief.

Backing work like Kenrex requires the same mindset. It means trusting artists, accepting uncertainty, and taking reputational and financial risk in pursuit of something new. That kind of risk-taking is increasingly rare, which is exactly why it matters.

My view is simple. Kenrex is brilliant. It is totally unique. It deserves a wide audience, and if you can see it, you should. I would be very surprised if this does not go on to an even bigger stage.

Although Kenrex is rooted in 1980s America, its concerns feel entirely current. The mechanism has not changed. Only the setting has.

Today, bullying rarely happens on a main street. It happens on a screen.

Cyberbullying and the Modern Fear Equation

My Yorkshire Post column this week focused on cyberbullying and the persistent belief that online cruelty is easier to ignore than its physical equivalent.

The advice sounds familiar. Do not engage. Rise above it. Do not feed the trolls.

As with Skidmore, that advice contains some truth. But it also avoids the harder reality. Online bullying persists for the same reason it always has. Because standing up to it carries risk.

Digital harassment is rarely a single incident. It accumulates. It repeats. It distorts. It follows people into their work, their families, and their sense of safety. The damage is quiet, but it is relentless.

What makes cyberbullying particularly effective is how it exploits fear in bystanders. Calling it out can trigger pile-ons, misrepresentation, or becoming the next target. So people choose the safer path. Private messages of support. Quiet sympathy. Public distance.

Most people do not join in. They simply decide not to get involved.

That decision is understandable. It is also decisive.

Silence creates the impression of consensus. Consensus emboldens behaviour. Over time, the bully feels untouchable, not because everyone agrees, but because fear has done its work.

The parallel with Kenrex is direct. Bullying survives not because communities endorse it, but because the personal cost of speaking up feels too high to carry alone.

Private support helps people endure. Collective, visible support is what changes behaviour.

A keyboard is not a shield. A platform is not a licence. Harm does not need bruises to be real.

Read The Column Here

Building Communities Where Fear Has Less Power

The final story this week looks forward.

This week marks the launch of the 2026 Digital Forge Membership. From next year, all Digital Forge events move to a members-only model.

That decision reflects a simple insight. It is easier to speak up when you are not doing it alone.

Events create energy. Communities create safety. Repeated interaction, shared norms and trust reduce the personal risk of saying, this is not acceptable. They turn individual courage into collective behaviour.

Digital Forge exists to bring together founders, engineers, operators and partners who are actively rebuilding British industrial capability. Membership allows us to go deeper. A coherent programme. A shared identity. A community that values contribution and accountability.

The 2026 programme reflects the real challenges facing the region. From Seed to Scale. Female Founders and the Capital Gap. AI and the New Era for Startups. Reindustrialising the North.

Members will also receive first access to live podcast recordings, workshops, trips and the inaugural Digital Forge Awards.

At its core, this is about building environments where fear has less power, and silence is no longer the default response.

To mark the launch, we are offering a 20 percent discount for early members using code FORGE20.

Join Now

Final Thought

Standing up to bullying is the right advice.

It is also hard advice to follow when fear isolates people.

Kenrex shows how fear allowed a bully to dominate a town.

My Yorkshire Post column shows how the same dynamic now plays out online.

The Digital Forge is about building communities where the cost of speaking up is shared.

Silence often feels safe.

It rarely is.

Read The Sunday Signal first every Friday on Substack.

Public release follows on Sunday.

The FCA has closed its investigation into WANdisco with no action taken. A two-year cloud has lifted. The noise that surrounded this period never reflected the truth. Now an independent regulator has confirmed it.

At the same time, I want to focus on what comes next. I spoke recently with Bruce Katz, one of the world’s leading thinkers on cities and industrial strategy, about the once-in-a-generation opportunity in front of South Yorkshire. And in my Yorkshire Post column this week, I explored how the cancellation of Football Manager 2025 reveals a much deeper crisis inside organisations that have ignored their own technical debt for too long.

Three signals. One message. You cannot build the future until you clear what sits unresolved behind you.

Subscribe now

Clearing the Record

When the FCA decision was published in Friday’s RNS, it finally put into writing what I had known from the start. No action. No criticism. No findings against me. I had acted properly, in good faith and with integrity.

But the hardest part of the past two years was not the process. It was everything that filled the silence around it.

From the moment I uncovered the irregularities in 2023 and escalated them, misinformation moved faster than truth. Entirely false briefings were handed to journalists by people who should have known better. Assumptions became stories. Stories became narratives. And those narratives were often written by people who had never been in the room and never had access to the facts.

It is a strange feeling to watch a version of your life being constructed by others. It is worse when you are not allowed to correct it in real time.

That noise caused real harm. Not just to me but to people around me who were already carrying their own pressures. My colleague and friend, Peter Scott, was dealing with the slow and painful loss of his father during the height of it.

There’s a line from Martin Luther King Jr that stayed with me:

“In the end, we will remember not the words of our enemies, but the silence of our friends.”

Those words stayed with me because they were true.

You learn who stands with you when the weather turns. Some people moved closer. Others drifted away. A few simply waited to see which way the wind would blow. Meanwhile, the quiet messages from those who believed in the truth landed with enormous weight. They mattered more than I can ever express.

That is why the FCA’s clarity this week carried such significance. An independent regulator examined every disclosure, every process and every decision. The conclusion is unambiguous. The concerns about my conduct were unfounded.

And then, on Tuesday, I saw something I will never forget.

A headline in the Yorkshire Post that read:

“Probe finds WANdisco’s ex boss acted properly.”

A simple sentence. Yet for my family, my colleagues and the people who supported me, it represented the end of a long and painful chapter. It meant that the truth had finally replaced the noise.

This experience has changed me. It has stripped away illusions about how public narratives form. It has sharpened my purpose. And it has reminded me that resilience is not a posture but a practice, developed quietly under pressure.

To those who stood by me, thank you.

To those who stayed silent, I heard that too.

A line has now been drawn.

Onwards.

A Once-in-a-Generation Opportunity for South Yorkshire

A few weeks ago I sat down with Bruce Katz to record an episode of The Digital Forge. His background is extraordinary. He helped craft US housing policy during the Reagan years, served as chief of staff at the Department of Housing and Urban Development under President Clinton, co-led the housing transition for President Obama and founded the Metropolitan Policy Program at Brookings. Today he runs the Nowak Metro Finance Lab and advises governments around the world.

I asked him what he sees when he looks at South Yorkshire.

His answer was crystal clear. The world has changed in ways that favour regions with deep industrial capability. Advanced energy. Advanced materials. Robotics. Secure supply chains. These are the arenas that now anchor global competition.

Bruce believes South Yorkshire has the assets the twenty-first century actually needs. Not nostalgia but real capability. Precision engineering. A world-class applied research ecosystem. The AMRC. A workforce that still understands how to design and build physical systems. Supply chains that already connect to nuclear technologies and could soon anchor Small Modular Reactor production.

He pointed to Pittsburgh as the nearest analogue. A city dismissed as a relic after its steel collapse, yet rebuilt as a global hub for robotics and AI. The turning point was not a federal decree. It was a local decision to build around its true identity.

Bruce’s advice to us was direct.

Stop waiting for permission from London.

Be honest about what this region is best at.

Unify institutions around that mission.

Tell the world what we do and invite capital to back it.

When I asked him what success would look like in ten years, his reply was striking.

“You would be globally recognised as one of the core nodes of advanced industry in the world.”

Not a slogan. A realistic outcome if we choose it.

The FCA decision has cleared the mental and emotional space for me to focus entirely on this mission. Yorkshire AI Labs exists because I believe South Yorkshire can lead Britain’s industrial revival. The opportunity is here. The question is whether we seize it.

Watch The Full Episode

When Old Software Sinks New Dreams

My column this week begins in a cold student house in the early nineties. Six of us. One PC. Championship Manager. Coloured dots drifting across the screen and hours disappearing into the early morning.

Three decades later, my son and his friends are doing the same with Football Manager. The dots returned, the rituals endured and the game became one of the most beloved franchises in the world.

Which is why this year’s announcement shocked so many. Football Manager 2025 was cancelled outright. Not delayed. Cancelled. The developers admitted that the new version was not good enough. Their transition to a new engine had gone wrong. Years of accumulated workarounds had piled up. Technical debt had caught up with them.

Technical debt is what happens when you choose the quick fix over the right fix. It is the old system nobody dares to replace. The dependency nobody understands. The patch upon patch upon patch that eventually collapses under its own weight. Every shortcut is a loan. The interest always arrives.

Sometimes it arrives brutally. Knight Capital lost hundreds of millions in under an hour because old code was accidentally reactivated. Equifax exposed the personal data of almost half the United States because a patch was not applied. Airlines have collapsed operationally because their scheduling systems were built for a world that no longer exists.

Football Manager’s decision was painful but brave. It hurt financially. It disappointed fans. But it preserved something more important than a yearly product cycle. It preserved trust.

There is a wider lesson here for leaders across every sector.

Britain is running on organisational and technical debt that can no longer be ignored.

Legacy IT. Outdated regulations. Business models designed for a different world.

If we want to compete in AI, if we want to reindustrialise, if we want to build the future, we must first repair the foundations. Football Manager faced its debt before it became a crisis. Most do not.

As for me, I have downloaded Football Manager 2026. I am attempting to take Buxton from the sixth tier to glory.

Final Thought

This week delivered clarity, opportunity and a reminder of the hidden debts that hold us back.

A regulatory chapter has closed.

A global expert has told us that South Yorkshire’s moment is here.

A cancelled football game has shown what happens when foundations are ignored for too long.

The common thread is simple.

You cannot build tomorrow if you refuse to confront yesterday.

Thank you for reading and for walking this journey with me.

Onwards.

Read The Sunday Signal first every Friday on Substack.

Public release follows on Sunday.

In August, I wrote that the mass lay-offs in Silicon Valley were not a correction but the opening act of an AI-driven restructuring of the labour market. This week proved the point. The pretence is over. Companies are now saying the quiet part clearly. AI is taking work that once belonged to people. Professions that built their identity on stability are being forced to confront their own disruption.

If you sell time, you are running out of it.

If you sell judgment, you are entering your strongest decade.

At the same time, Britain’s ambition to lead in AI collided with the reality of its energy system. For all the talk of innovation, nothing moves without electricity. And we are nowhere near ready.

And then, in an altogether different way, Paul Gascoigne reminded us that AI still struggles to understand large parts of the society it claims to serve. Technology may speak in universal terms, but it still listens through a narrow set of ears.

Three signals. One truth. AI is not a distant horizon. It is the present tense.

Subscribe now

The Jobs Shock Has Begun

In that August column, I argued that the first unmistakable sign of the AI labour shock would not appear in think tank reports or political speeches. It would appear in payrolls. The moment companies stopped hiding behind euphemism and said plainly that machines were doing the work people once did. That moment arrived this week.

HP announced plans to cut thousands of jobs as it embeds AI more deeply across its operations. That reduction represents roughly ten per cent of its global workforce. McKinsey, the consulting giant that once guided others through disruption, has cut more than ten per cent of its own headcount over the past eighteen months. Allianz is preparing to remove up to eighteen hundred call centre roles as AI systems take over routine customer service and claims handling. Clifford Chance, one of London’s leading law firms, has cut business services posts and reshaped others as technology takes on more internal work.

These are not hygiene cuts. They are strategic resets.

More striking still is the shift inside the legal profession itself. Only a year ago, most major firms believed AI would lift revenue and improve margins. That confidence has evaporated. Today, a clear majority expect AI to put downward pressure on both. More than half expect clients to demand the same work at lower cost because they know machines are doing a growing share of it. Others expect clients to bring more work in-house as they adopt AI tools directly. Firms now anticipate a double-digit fall in chargeable hours as automation becomes standard.

For decades, the professional services model relied on a pyramid. Partners at the top, juniors at the base, all held together by process. Partners sold trust and experience. Juniors supplied the hours. But AI now does the work that once filled those hours. The pyramid is flattening.

In August, I wrote that if even software engineers could no longer rely on job security, then no one should assume safety. That warning looks modest now. The roles disappearing today are not low-skilled or peripheral. They are the foundational jobs that shape entire careers.

Automation does not nibble at the edges. It clears space. It replaces the predictable and rewards the exceptional. It punishes those who rely on repetition and elevates those who can interpret, create and decide.

The question now is not whether AI will reshape the labour market. It already has. The question is whether Britain is prepared to reskill, retrain and rethink fast enough to keep up.

Britain’s Power Problem: The Hidden AI Bottleneck

While job losses dominated the headlines, another constraint emerged that may ultimately decide whether Britain competes or falls behind. AI does not run on optimism. It runs on electricity. And the numbers are staggering.

In my Yorkshire Post column on Friday, I wrote that the government’s current goal of six gigawatts of AI-grade data centre capacity by 2030 already understates the true need. Six gigawatts is roughly the electricity demand of every home in Yorkshire.

Yet developers have already announced between 500 and 900 megawatts of new capacity that will need to be powered in the next few years alone. One operator is seeking an entire one-gigawatt site dedicated specifically to AI. And, most striking of all, National Grid is processing up to nineteen gigawatts of new connection requests from data centre operators for delivery by 2031.

Nineteen gigawatts is the electricity consumption of roughly fourteen million homes. A new industrial sector is landing on the grid in a single decade, and we are entirely unprepared for it.

The consequences are already visible elsewhere. In California, brand-new data centres stand idle because local utilities cannot deliver the power. Some of the world’s most advanced GPU clusters are sitting on warehouse floors because they cannot be plugged in. The bottleneck is no longer the chip. It is the socket.

China has taken the opposite approach. It is scaling nuclear, hydro and thermal baseload at a pace the West no longer attempts. AI leadership follows energy leadership. The country that builds the megawatts will lead the models.

This is why South Yorkshire’s vision matters so profoundly. At the Digital Forge Summit, Richard Caborn argued that the region should anchor Britain’s clean power future by becoming the manufacturing base for Small Modular Reactors. The proposed National SMR Manufacturing Centre of Excellence already has more than fifty industrial partners and private customers ready to place orders for British-built reactors.

It is the most advanced SMR proposal in the country. If the first reactor is built abroad, the supply chain follows it. If it is built in Yorkshire, this region becomes the powerhouse of Britain’s AI economy.

AI is not constrained by imagination. It is constrained by electricity. And Britain must decide whether it intends to be a leader or a spectator.

Read My Yorkshire Post Column

What Gazza Taught Silicon Valley

The most surprising story of the week did not come from a boardroom or a data centre. It came from a recording booth.

Victoria Williams, the ghostwriter of Paul Gascoigne’s new autobiography, tried to use AI tools to transcribe their interviews. The software failed. Completely. It could not understand his Geordie accent. Words like “hide” became “hate”. “Lies” became “lawyers”. Entire passages were lost. In the end she transcribed every word by hand to avoid putting phrases into his mouth that he never said.

On the surface it is amusing. AI cannot understand Gazza. But beneath the humour is a flaw that runs through the foundations of the industry.

Speech recognition systems were trained on what was easy to collect: standard American English, certain southern English accents and the speech patterns of a narrow socioeconomic group. Huge parts of Britain were never included. Accents from the North East, Scotland, Wales and Northern Ireland routinely trip up commercial models. Many Black British and South Asian voices are misinterpreted or ignored entirely.

We have seen this pattern before. Facial recognition systems that worked well on lighter skin and poorly on darker skin. Hiring tools that downgraded CVs because they did not match historical patterns. The issue was not ideology. It was data. Technology reflects the sample it is trained on, not the society it is deployed into.

Gazza exposed the flaw by accident. The machine did not understand him because the machine had never learned to.

Williams said the manual transcription gave her a deeper understanding of his voice and his story. That is the human advantage. But we cannot rely on that advantage alone. When AI is used to interpret witness statements, describe symptoms, assess loan applications or respond to children in classrooms, the cost of mishearing becomes serious.

If AI cannot hear everyone equally, it cannot treat everyone equally. And if it cannot treat everyone equally, it cannot serve the society it claims to transform.

Gazza, without meaning to, demonstrated that AI still carries the biases of the world that built it. The next stage of the AI revolution will not be measured only by speed or power but by inclusion. A technology that understands only part of the population cannot claim to shape the whole of the future.

Final Thought

This week delivered three warnings and one opportunity. AI is reshaping the labour market faster than governments can respond. It is colliding with an energy system that was not designed for it. It is revealing biases that have been ignored for too long.

Yet inside these disruptions sits the chance to rebuild. Regions that choose to supply the energy backbone of the AI age will lead. Workers who embrace AI as an amplifier rather than a rival will thrive. Countries that demand technology built for every voice will become the places where innovation actually works.

The future will not wait for those who want more time. It will reward those who act.

Read The Sunday Signal first every Friday on Substack.

Public release follows on Sunday.