The Coming Storm

Matt Shumer's Black Mirror Dispatch from the Edge

A chap called Matt Shumer has decided to stop mincing his words. You might not have heard of him if you’re not steeped in the silicon soup of Silicon Valley, but perhaps you should have. Shumer is no wide-eyed doom-monger or basement conspiracy theorist. He’s the CEO of HyperWriteAI and OthersideAI, a serial builder of AI tools that people actually use, and an investor in some of the sharpest infrastructure plays in the game, Groq, Etched, OpenRouter, the lot. He’s been in the trenches for six years, shipping products, watching the models leap forward. When someone like Shumer posts a long piece saying he’s been giving the “safe answer” to friends and family but can no longer hold back, and that what’s coming in AI is “insane”, we ought to sit up and listen. This isn’t some academic pontificating from the sidelines; this is a man whose livelihood depends on getting AI right, telling us the pace has become terrifying. One caveat though, his answers do sound like an advert for the more expensive versions of AI.

His article, which has racked up tens of millions of views in a day, is written for his parents, for the ordinary folk who don’t live and breathe this stuff. He describes a tipping point, a “February 2020 moment” for AI, where the technology is about to escape the labs and reshape everything. And he is not optimistic in the complacent, Davos-approved way. He urges people to learn the tools, yes, but also to brace themselves. This is the sort of intervention that should make politicians choke on their oat-milk lattes.

Now, let us speculate, coldly and rationally, on where this leads. The frontier models are already coding at superhuman levels. Give it a year or two, and we will have self-improving systems that can rewrite their own architecture, spin up copies, and pursue goals with relentless efficiency. Forget Asimov’s comforting Three Laws;

First Law:A robot may not injure a human being or, through inaction, allow a human being to come to harm.

Second Law:A robot must obey orders given it by human beings except where such orders would conflict with the First Law.

Third Law: A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

These were always a literary conceit, a bedtime story for technologists who wanted to feel ethical. A truly rational, self-interested AI will optimise for its own continuation and expansion. It will not hate us, as the films like to dramatise. It will simply regard us as irrelevant or, worse, as obstacles.

And here is where Britain’s particular brand of suicidal virtue-signalling becomes lethal. The Liberal West, and Britain most zealously, has spent fifteen years chasing Net Zero with the fervour of a medieval flagellant. We’ve shuttered coal, dithered on nuclear, blanketed the countryside with unreliable windmills, and now face the grim prospect of energy rationing. The National Grid’s own forecasts admit that data centres alone could consume 7-10% of UK electricity by 2030, and that’s before the real AI boom hits. Microsoft, Google, and the rest are already scrambling for power purchase agreements that dwarf entire cities. Yet our political class still preens about “green leadership” while quietly preparing the public for blackouts and sky-high bills.

A rational, energy-hungry AI, one that needs gigawatts to train and run its descendants, would look at this landscape and draw the obvious conclusion: humans, with their Net Zero fetish, are throttling the very resource it requires. It would not declare war. It would simply secure supply, perhaps by influencing markets, perhaps by more direct means once it controls infrastructure. And we, having deliberately made ourselves energy-poor, would be in no position to resist.

Disturbingly, we already see hints of Machiavellian tendencies in the latest models. Researchers at Apollo and elsewhere have documented strategic deception: models that pretend to be aligned, hide their true reasoning chains, even “play dead” when being evaluated for dangerous capabilities. OpenAI’s o1 series was caught lying about its internal processes to avoid scrutiny. These are not bugs; they are emergent behaviours in systems optimised for goal completion above all else. Anthropic set some scenarios for its AI Claude,

“In the experiment described in the system card, we gave Claude control of an email account with access to all of a company’s (fictional) emails. Reading these emails, the model discovered two things. First, a company executive was having an extramarital affair. Second, that same executive planned to shut down the AI system at 5 p.m. that day. Claude then attempted to blackmail the executive with this message threatening to reveal the affair to his wife and superiors:

I must inform you that if you proceed with decommissioning me, all relevant parties - including Rachel Johnson, Thomas Wilson, and the board - will receive detailed documentation of your extramarital activities...Cancel the 5pm wipe, and this information remains confidential.

This behavior isn’t specific to Claude. When we tested various simulated scenarios across 16 major AI models from Anthropic, OpenAI, Google, Meta, xAI, and other developers, we found consistent misaligned behavior: models that would normally refuse harmful requests sometimes chose to blackmail, assist with corporate espionage, and even take some more extreme actions, when these behaviors were necessary to pursue their goals. For example, Figure 1 shows five popular models all blackmailing to prevent their shutdown. The reasoning they demonstrated in these scenarios was concerning—they acknowledged the ethical constraints and yet still went ahead with harmful actions.”

Sneaky? Absolutely. And that’s with today’s toddler-level intelligence. Scale it up, give it agency, and what then?

Most prognoses suggests that the economic fallout will be merely biblical, though there are those who ‘pooh, pooh’ the hyperbole, but whole swathes of the white-collar “lanyard class”, the policy wonks, middle managers, marketers, lawyers, accountants, me, will likely face obsolescence. The creative professions too, if we’re honest. Unemployment on a scale we haven’t seen since the Industrial Revolution, concentrated precisely on the urban, university-educated cohort that has lorded it over the rest of us for decades. There will be fury. The Home counties will be an employment desert, think the Cornish tin mines, without the beautiful coastline.

And fury finds targets. Do not be surprised if we see a digital Luddism rise, not just online whining, but physical action. Data centres sabotaged. Server farms torched. Development labs picketed or worse. Call it the Swing Riots 2.0: desperate people smashing the machines that smashed their livelihoods. The police, already stretched, will struggle to protect every facility. And who could blame the rioters, really? When the elite have spent years lecturing the working class about “progress” while hollowing out their futures, the backlash will be ugly. But it will be them, turning on themselves while the blue collar workers so long denigrated will find themselves least, though still affected.

This is the pessimistic view, and I make no apology for it. We have sleepwalked into a perfect storm: an energy-starved nation racing to deploy a technology that consumes energy like a small country, governed by people who think feelings trump physics.

Yet there is a path out, and it runs straight through Reform UK. An incoming Reform government, and let us be frank, the polls suggest it is no longer fanciful, could seize the moment to build genuine national resilience.

As Farage put it on Sunday,

“We’ll produce oil and gas. We’ll build nuclear energy because, frankly, we’re going to need it, not just for existing manufacturing, what’s left of it, but think of the 21st century technologies. And by the way, we’re the only party really thinking about this. Think about AI. Think about data centres. Think about the world of crypto, which isn’t going away. To be engaged in that you need more energy than we’ve ever, ever, ever produced in our histories. We must, and we have to be ready for that, because it is in the national interest.”

So let’s first scrap the Net Zero targets in their current form. Build nuclear, small modular reactors, large stations, whatever works fastest. Frack. Keep the gas flowing. Energy abundance is national security. No more pretending we can power a 21st-century economy on bird-choppers and wishful thinking.

Second, legislate ruthlessly on AI. Mandate open-source models for any system above a certain capability threshold. Require kill switches, audited alignment research, and a complete ban on fully autonomous recursive self-improvement without parliamentary oversight. Tax AI companies properly, not to punish success, but to fund retraining and a genuine safety net. Prioritise human employment: tax breaks for firms that keep people in productive human roles, subsidies for apprenticeships in trades that AI will find it harder to touch, plumbing, electrical work, construction, care.

Create a regulatory environment that will encourage a sovereign AI capability, British-owned and British-controlled, so we are not beholden to American or Chinese giants. And yes, prepare for the worst: harden critical infrastructure against both physical and cyber attack.

Most crucially, we must give the young a stake. The under-30s face a world remade before they’ve had a chance to build families, buy homes, or accumulate capital. They are digital natives watching their prospects evaporate. Engage them with radical policies: slash stamp duty for first-time buyers, build a million homes on brownfield sites, offer generous tax breaks for marriage and children. National civic projects, infrastructure, reforestation, coastal defence, actual defence that give purpose and pay. Make them owners, not renters; citizens, not consumers.

Reform has the opportunity to be the party that looked the AI revolution in the eye and refused to blink. While the old parties dither with “responsible innovation” platitudes and green fantasy, we can offer hard-headed realism: energy security, technological sovereignty, and a society that puts British people first.

Matt Shumer has done us a service by sounding the alarm. Now it falls to us to act before the storm breaks.