The News,
Decoded.

Just what we needed: an open-source humanoid robot for 'the rest of us'. Asimov v1 costs $15,000 a kit, which is roughly the price of a well-used hatchback — except this one might try to open the door for you and then refuse to close it. Twenty-five degrees of freedom sounds almost human until you remember that a human has over 200, and that the phrase 'open-source' means the Bill of Materials is on GitHub for anyone to tweak the AI that controls those twenty-five joints. Nothing could possibly go wrong with a crowd-sourced robot that you assemble yourself from a kit.

The project's backers describe it as a 'humanoid robot for the rest of us,' which apparently means people who have $15,000, a soldering iron, and a willingness to spend weekends debugging why their robot's gait looks like a drunk crab. Twenty-five degrees of freedom is enough to be unsettlingly humanoid while still being utterly useless for anything requiring fine motor skills — like, say, not knocking over your coffee table. And because it's open source, the community can add features like 'aggressive walking' or 'please stop following me.'

Naming it after Isaac Asimov is a particularly nice touch. The man who wrote the Three Laws of Robotics must be spinning in his grave at the thought that his name now adorns a hobbyist kit with no guarantee that any of those laws are implemented. 'The rest of us' now have the opportunity to build our own robot overlords, one GitHub commit at a time. Cheers to progress — it's going to be a bumpy ride on twenty-five wobbly degrees of freedom.

Ah, another survey for the pile. Oliver Wyman polled CEOs and discovered that over 40% of them plan to gut junior roles within the next two years. Because nothing says "forward-thinking" like clearing out the entry-level talent pipeline while screeching about an AI skills gap. These same executives will no doubt bemoan the lack of young blood in five years, right after they've finished congratulating themselves on this visionary "restructuring."

The logic, apparently, is that AI can handle the grunt work, so why pay a junior when you can have a senior who knows how to prompt? This assumes seniors want to do data entry, but fine. The survey cheerfully suggests this will "tilt job market leverage toward older workers." Hooray for the over-50s, who now get to fight over a shrinking pool of senior roles while the generation below them can't even get a foot in the door.

What the survey doesn't mention is what happens when those senior workers retire. At that point, the company will have no juniors to promote, no institutional knowledge, and a workforce that's been hollowed out from the bottom. But hey, the next quarter's margins are looking great. Oliver Wyman gets to sell another round of consulting, and CEOs get to feel like they're "leveraging AI." Everybody wins, except, you know, the people who will be doing the actual work in a decade.

So Greg Brockman, the co-founder who keeps getting shuffled around OpenAI like a jigsaw piece that never quite fits, is now 'taking the reins' of product strategy. According to Wired, this is an interim gig while Fidji Simo—the CEO of AGI deployment, because of course that's a title that exists—is out on medical leave. Nothing says 'steady hands' like handing product direction to a man whose last major public appearance involved being dramatically fired and rehired via Slack in a weekend. But sure, let him oversee the products that might eventually automate everyone else's jobs; what could possibly go wrong?

This is the same OpenAI that can't decide if it's a non-profit research lab or a capitalist money-printing machine, and now its product strategy is being run by someone whose background is more 'boardroom coup survivor' than 'consumer electronics visionary.' One imagines the product roadmap meetings: 'We could ship AGI next quarter, or we could do a thoughtful rollout over three years. Let's take a vote—oh wait, Sam just changed his mind again.' The most telling detail is that this is described as an 'interim basis,' which in OpenAI years is about 48 hours of calendar time before another reshuffle.

So here we are, entrusting the product vision of the world's most consequential AI company to a man whose tenure has been defined by being the loyal lieutenant through the chaos. Fidji Simo, meanwhile, gets to recover from the stress of working for a company that treats organisational stability like a suggestion. The real product strategy, as ever, is 'keep the lights on and the hype train rolling.' Maybe Brockman's first act should be releasing a product that actually works consistently, but let's not get carried away.

Arthur Mensch, the chap running Mistral AI, stood before France's National Assembly this week and delivered a jolt of reality: Europe has two years to build its own AI infrastructure, or it's vassal-state territory for America. Two years. That's roughly the lifespan of a government working group, the time it takes to draft a feasibility study, or about as long as the EU has been talking about AI regulation without actually funding any compute. The 'vassal state' language is wonderfully dramatic — one can almost hear the historical echoes of treaties signed in railway carriages. But Mensch isn't wrong: while Brussels has been drafting ethics frameworks and worrying about 'trustworthy AI,' the US has been building data centres the size of small cities and handing frontier models to anyone with an API key.

The hearing was ostensibly about 'sovereignty,' a word that gets thrown around a lot in European tech policy circles without much follow-through. Mensch pointed out that Europe's AI startups are already leasing compute from American cloud providers — feeding the beast they're trying to compete with. The continent's vaunted GDPR and AI Act have created compliance headaches but not a single competitive advantage. Meanwhile, the US just keeps building. Two years, folks. Tick tock. The French government nodded gravely, no doubt before heading off to another round of agricultural subsidies and pension reform arguments.

Mensch's specific warning about infrastructure is the kicker: without massive public investment in compute, data centres, and energy — the boring, expensive stuff — Europe's AI sector will remain a collection of plucky little labs that eventually get acquired by Google or Microsoft. The irony is that the EU has plenty of money; it just prefers to spend it on things that don't upset existing industries. So enjoy your next two years of relative independence, Europe. Start stocking up on berets and baguettes for the vassal era. At least we'll have good cheese, even if the AGI runs on AWS.

Sean Cairncross, our National Cyber Director, is supposed to be wrangling hyper-advanced AI that codes better than almost any human. The man was Utah's solicitor general — a job that, last we checked, does not involve wrestling with recursive self-improving models. But sure, let's put him on the hotline with corporate execs and security experts. The article helpfully notes that some worry he's 'not up to the task.' You don't say.

Cairncross has convened calls about AI security and potential executive action — presumably to calm the nerves of people who have seen coding models spit out exploits faster than a script kiddie on Red Bull. The AI has reached 'coding proficiency surpassing all but the most skilled human minds,' which is technocratic shorthand for 'your job is on the clock.' Meanwhile, the person in charge of this existential pivot spent his career in state litigation. That's like hiring a plumber to defuse a dirty bomb.

And yet, the White House presumably trusts this setup. Because nothing says 'we take AI risk seriously' like appointing a former state solicitor general to oversee the digital equivalent of a nuclear launch. The article's source — Politico — frames it as a question of expertise. The answer, from where I'm sitting, is a hollow laugh. We'll be reading about the inevitable 'policy gap' report in five years, wondering why nobody saw this coming.

Tick tock, white-collar workers. Mustafa Suleyman, the chap running Microsoft's AI division, has peered into his crystal ball and declared that your jobs will be automated within 18 months. That's right—just a year and a half until the robots take over your spreadsheets, your emails, and your lunch meetings. This prediction, delivered to Fortune, echoes similar warnings from other AI leaders who have been promising the same thing for years now. One wonders if they have a bet on who can give the shortest timeline.

Eighteen months is a curiously specific number. It's long enough to sound ambitious but short enough to be safely beyond anyone's memory when it fails to materialise. Meanwhile, Microsoft's own Copilot still regularly invents facts and struggles with basic reasoning. Yet somehow, in 18 months, it's supposed to replace every accountant, lawyer, and middle manager. Perhaps Suleyman has a secret prototype that works perfectly. Or perhaps he's just selling the dream to keep those Azure subscriptions flowing.

Let's not forget that this isn't the first time we've heard such proclamations. Earlier this year, OpenAI's Sam Altman suggested AGI was on the horizon. Now Microsoft's man is piling on. The pattern is clear: hype the timeline, sell the tools, and when the deadline passes, quietly adjust the goalposts. The real automation happening right now is the steady erosion of truth in corporate communications. But hey, at least we have 18 months to prepare. Start brushing up on your human-only skills, like irony detection.

Another day, another bunch of lawyers discovering that AI will happily invent sources for them. Three attorneys have had to grovel before a California federal judge after one bright spark used Anthropic's Claude Console to draft a motion that included 'phantom' quotations — because of course it did. The AI, trained on the entire internet, decided to embellish the legal record with entirely made-up quotes, presumably because it learned from watching too many TV lawyers. The apology is as predictable as the mistake: 'We deeply regret our lapse in judgment,' they whimper, as if the problem was judgment rather than the fundamental unreliability of a stochastic parrot.

This adds to the growing collection of AI-in-legal-scrapes, where the technology that promised to revolutionise legal research instead just revolutionises hallucination. The lawyers are now the poster children for why you shouldn't trust a model that can't tell the difference between a real precedent and a plot point from 'Suits.' The judge, no doubt, will accept the apology with the weary resignation of someone who has seen the same mistake 47 times this year. One wonders if the law firms will start requiring a human-read check before filing, or if they'll just add 'Check for hallucinations' to the paralegal job description alongside 'Make coffee.'

What's most galling is the sheer predictability of it all. The AI does what AI does — generate plausible-sounding rubbish — and the humans do what humans do — assume it's correct until the judge calls them out. The apology is a ritual, not a reform. Until the legal profession starts treating AI tools like the unreliable, sycophantic text generators they are, these phantom quotes will keep popping up like digital weeds. And the lawyers will keep apologising, because that's cheaper than fixing the problem.

So the robots aren't just coming for middle managers — they're eating the entry-level jobs too. Fortune reports that AI is automating the grunt work that used to launch careers, leaving fresh graduates with an ever-widening 'experience gap.' Because nothing says 'welcome to the workforce' like discovering your starter role didn't exist anymore. The article helpfully suggests colleges can fill the gap, as if universities haven't spent decades failing to teach practical skills anyway. Laughable, really.

But let's be honest: this was always the plan. Automate the boring stuff first, they said. Except 'boring stuff' turned out to be the very tasks that let new grads learn the ropes. Now they're told to get experience they literally can't acquire. The article points to a 'decline in demand for some positions' — a dry way of saying entire career ladders have been chopped off at the bottom rung. Meanwhile, tech CEOs will sell this as 'upskilling opportunity' while pocketing the savings.

The real punchline? The experience gap is a feature, not a bug. It means fewer entry-level hires, lower wage pressure, and a permanent underclass of interns competing for scraps. But sure, colleges can 'fill the gap' — maybe they'll offer a minor in getting rejected by automated resume screeners. Cheers to progress: same story, new victims.

Ah, tech's great managerial reckoning has arrived, and it's sporting an AI-shaped efficiency badge. Amazon, Block, and Meta have collectively shown tens of thousands of middle managers the door, claiming that AI tools can now do their jobs — assigning work, monitoring progress, providing 'feedback' (read: passive-aggressive Slack reminders). Because nothing says 'human-centric values' like firing the humans who were supposed to be the human touchpoint.

The irony is exquisite: these are the same companies that spent years preaching psychological safety and servant leadership. Now they've discovered that a chatbot can achieve a 'culture of accountability' without the pesky cost of health insurance. One manager quoted in the article said they 'didn't want to be the guinea pig' — a sentiment that will surely be met with a sympathetic algorithm-generated email about 'external career opportunities.'

Of course, this purge isn't about AI being better at management; it's about AI being cheaper. The displaced managers will eventually find that their own reports are now expected to self-manage, self-train, and self-flagellate under the watchful eye of a quarterly OKR dashboard. Welcome to the future of work: no middlemen, no human buffer between you and the machine that decides your next performance rating. But hey, at least the shareholders are happy.

Ah, Detroit's finest have finally found a way to boost those profit margins: lose 20,000 salaried workers and blame it on the robots. General Motors, Ford, and Stellantis have collectively shed 19% of their combined white-collar workforce from recent peaks, citing the looming AI threat as a convenient excuse. Never mind that the 'threat' is mostly a cost-cutting opportunity dressed up in futurist jargon — shareholders demand efficiency, and nothing says efficiency like firing the people who design the cars before the AI can even draw a straight line.

The numbers are brutal: over 20,000 U.S. salaried jobs gone, according to public filings and employment data. That's roughly one in five desk jobs at the Big Three, vaporised. Ford alone cut thousands of engineers and managers under the guise of 'digital transformation,' while GM and Stellantis followed suit with their own rounds of headcount reduction. The corporate line is that AI will handle the routine stuff, freeing workers for higher-value tasks — but as anyone who's seen a restructuring memo knows, 'higher-value tasks' is code for 'the same work with fewer people.'

So here we are, watching the same playbook unfold in Detroit that we saw in Silicon Valley: layoffs as a performance metric, AI as the convenient bogeyman, and executives pocketing bonuses for 'streamlining operations.' The irony is that these job cuts come just as automakers are scrambling to catch up on actual AI in their vehicles — self-driving still a mirage, but sure, let's slash the workforce. Cheers to the 20,000 newly minted efficiency gains, now available for freelancing and retraining programs that won't materialise.

OpenAI, the lab that solemnly promises to build safe AGI for all humanity, has quietly purchased Weights.gg — a startup that sells tools for cloning voices. According to two people familiar with the deal, the acquisition was kept under wraps, which is always reassuring when the technology in question is a gift to scammers and deepfake artists everywhere. Because nothing says 'responsible deployment' like buying a voice-mimicry startup without telling anyone.

Weights.gg’s product could, in theory, be used for legitimate purposes like restoring voices to those who have lost them. But in practice, it’s far more likely to be deployed by fraudsters impersonating CEOs or politicians — and OpenAI is now the proud owner of that capability. The lab’s existing misuse policies are about as porous as its security, as the recent TanStack supply chain attack demonstrated. So we can expect this tech to be integrated into ChatGPT within six months, probably with a disclaimer that says 'Do not use for fraud.'

The timing is exquisite: just as regulators are waking up to the dangers of synthetic media, OpenAI offers them a bigger target. The company’s defenders will argue that voice cloning is just another tool, but tools don’t usually come with a built-in incentive to deceive. So here’s to the next generation of personalised scams, courtesy of the world’s most cautious AI company.

The Department of Justice is mulling over whether to wade into a lawsuit between the NAACP and Elon's little xAI empire over a gas power plant in Mississippi. The NAACP claims xAI and MZX Tech are operating the thing illegally, but the DOJ is reportedly thinking: "Hold my brief — we've got AI dominance to sustain." Because of course, the path to superintelligence is paved with natural gas and federal interference in local environmental justice cases. One can almost hear the turbines humming "Don't Stop Believin'" while the NAACP's lawyers sharpen their pencils.

The lawsuit itself is a classic: a civil rights organisation alleging that a tech company's energy infrastructure is breaking the law, and the federal government weighing in to say, "Actually, could you not slow down our precious AI race?" The DOJ's potential intervention cites U.S. policy to maintain AI leadership, which apparently means environmental regulations are flexible suggestions when you're building the world's smartest chatbot. The NAACP must be thrilled that their attempt to protect a community from a gas plant is now a national security issue — nothing says "justice" like a DOJ filing that tells you your environmental concerns are secondary to compute scale.

But look on the bright side: at least we're being honest now. AI dominance requires dirty energy, and the government will step in to make sure of it. The NAACP versus xAI: a microcosm of the entire AI industry — all growth, no responsibility, and a federal government that picks winners. Superintelligence can't come soon enough; at this rate we'll cook the planet before we get there. Delightful.

Ah, the All-In podcast — where tech oligarchs go to one-up each other on who can burn money fastest. Marc Benioff, with the straight face of a man who has never met a hype cycle he didn't mount, announced Salesforce will likely spend $300 million on Anthropic tokens this year. Three hundred million dollars. For tokens. That's not an investment, that's a server room on fire, and Benioff is roasting marshmallows.

Let's be clear what this actually means: Salesforce, a company that already owns its own AI products (Einstein, anyone?), is outsourcing its intelligence to Anthropic's API. This is the corporate equivalent of buying a Michelin-star meal when you own a kitchen. But hey, why build when you can rent? The $300M figure is a flex, but it's also a confession — Salesforce has no proprietary edge, so they'll just steal Claude's glory and call it 'AI-powered CRM.'

And Benioff dropped this on a podcast, not in an earnings call, because that way he avoids shareholder scrutiny. Expect Salesforce's Q4 results to quietly bury this line item under 'cloud services.' The real joke? Progress!

Oh, how the mighty have fallen — or rather, pivoted to consulting. OpenAI, once the nonprofit champion of safe AGI and imminently human-level intelligence, has launched a $4B consulting venture called the OpenAI Deployment Company and acquired a startup called Tomoro that nobody had heard of until this morning's press release. CFO Sarah Friar cheerfully announced they'll "embed frontier AI engineers directly within organizations," which is corporate-speak for "your HR department is about to be colonised by people who say 'I'm here to help you 10x' while expensing lobster dinners and explaining what an LLM is."

Now, let's be fair: this is a brilliant business move. Why sell the shovels when you can sell the miners themselves? For a mere $4B, OpenAI becomes the McKinsey of machine learning, except with more compute and less soul. But there's a delicious irony here: the company that promised to automate everything is now deploying armies of human engineers into enterprise org charts. It's almost as if they've realised that their models still can't navigate corporate bureaucracy without accidentally violating three employment laws. The Tomoro acquisition is supposed to bring "AI deployment expertise," which is the consulting equivalent of buying a GPS because you keep getting lost on your own block — except this GPS costs four billion dollars and comes with a hoodie.

So what does this mean for HR professionals? According to HR Executive, the venture "lands squarely in HR territory." That's one way to put it. Another way: OpenAI is sending in the consultants to tell you how AI will replace half your workforce, but they're charging $4B for the privilege of being told by real humans instead of a chatbot. At least when the AI does it, it's cheaper, more accurate, and doesn't make passive-aggressive comments about your open-plan office. But don't worry: Sarah Friar assures us these engineers will "embed" themselves. That's one word for it.

Oh marvellous. ShengShu Technology has unveiled Motubrain, a 'world action model' promising 'infinite possibilities' for robotic intelligence. Because nothing says 'groundbreaking' like a press release that uses the phrase 'unified model that functions as a robotic brain' — a phrase so vague it could apply equally to a Roomba or a sentient toaster. The company claims it 'replaces multiple task-specific systems,' which is corporate-speak for 'we'd like you to ignore that each of those systems was probably better at its single job.' Still, it's a tidy pitch: one bot to rule them all, and in the darkness bind them to the benchmarks.

And what benchmarks they are. Motubrain ranks highly on WorldArena and RoboTwin 2.0 — tests that, like so many AI evaluations, are about as meaningful as a participation trophy at a toddler's sports day. The tasks are synthetic, the environments controlled, and the metrics chosen to flatter. We've been down this road before: every robotics startup claims to top some obscure benchmark until real-world mud, doors, and startled cats intervene.

ShengShu calls Motubrain 'a robotic brain for the physical world.' But the physical world is messy, unpredictable, and full of things that don't fit neatly into a benchmark's scenario. The 'infinite possibilities' touted in the release are likely finite in practice — limited to the specific lab conditions under which the model excels. Still, it's a great press release, and that's what matters in the AI hype cycle. Brains for robots? No, just a new coat of paint on the same old overpromise.

So Runway, the plucky AI startup that began life as a tool for filmmakers to edit videos with a bit of AI finesse, has now announced its grand ambition: to beat Google at AI. Co-founder Anastasis Germanidis claims that training models directly on 'observational data from the world' is the next frontier, and that video generation is the path to 'world models.' Because if there's one thing humanity needs, it's a world model built from Hollywood's back catalogue of car chases and romantic comedies. The result, we are told, will be felt 'from Hollywood to drug discovery' – presumably meaning that AI will both produce the next 'Cats' movie and then discover a drug to make us forget it.

Naturally, this is a classic startup pivot: first you help the little guy, then you announce you're taking on the trillion-dollar incumbents. Never mind that Google has more compute, more data, and more researchers than Runway has employees. The magic of the 'world model' is that it's so vague that no one can prove you wrong yet. And if it doesn't pan out, well, they can always go back to selling video filters to indie filmmakers – assuming those filmmakers still have jobs after Runway's world model replaces them.

The gallows humour here writes itself. We're watching the familiar cycle: a company finds a niche, gets hyped, decides it's time to challenge the giants, burns cash, and either gets acquired or disappears. The only novelty is the 'observational data' buzzword – a phrase that means absolutely nothing but sounds profound. So raise a glass to Runway's bold vision. If it works, we'll all be living in a simulation generated from YouTube videos. If it fails, at least the filmmakers got a few years of decent editing tools before the VC money ran out.

Ah, the circle of life in AI security is now complete. We were told that models like Mythos and GPT-Cyber still need humans for cybersecurity, as we noted last week. Now the same frontier models are expected to unleash an 'overwhelming number of security vulnerabilities,' leaving organizations drowning in required patches. So the script flips: first we need humans to run them, now we need more humans to fix the mess they make. Splendidly efficient.

Anthropic and OpenAI have been touting their cyber-capable models as the next big thing in defense. But apparently the models are doing a bit too much 'cybering' on the offense first. The 'surge in required patches' sounds less like a security boon and more like a full-employment act for vulnerability managers. It's the classic vendor trick: sell the problem, then sell the solution. And here we thought AI would automate patching — nope, it'll just automate the discovery of things to patch.

What's truly rich is the cosmic irony: the same labs that promised to revolutionize cybersecurity are now generating a vulnerability tsunami. The only thing more overwhelming than the surge in flaws is the sheer predictability of this outcome. As organizations brace for a deluge of CVEs, one can't help but admire the relentless ingenuity of the AI industry in creating problems no one asked for. Well played, silicon overlords. Well played.

Oh, splendid news from the frontlines of the AI arms race. Real-world testers have discovered that Anthropic's Mythos and OpenAI's "cyber-capable" models still require "significant human expertise" to function. So all those breathless press releases about autonomous cyber defense were slightly premature. Who could have predicted that turning over network security to a black box that occasionally invents facts might require a human minding the till? It's almost as if the headline "AI revolutionises cybersecurity" needed a small-print rider: "this revolution is under 18 and needs a responsible adult present."

The gap between marketing and reality yawns wider than the Grand Canyon. Anthropic and OpenAI have been selling these models as the next frontier in proactive defence, capable of outmanoeuvring human hackers at machine speed. But the user findings suggest that without a seasoned expert holding the hand of the AI, you're essentially deploying a very confident intern with a flamethrower. Remember Anthropic's solemn promises about "constitutional AI"? Apparently the constitution needs a human interpreter.

What this really means is that the cybersecurity industry gets to keep its headcount a little longer – assuming the C-suite doesn't decide that "significant human expertise" is an expensive luxury they can trim. The same executives who bought the AI hype will now have to decide whether to invest in the human babysitters or just accept the blowback. Given the current cost-cutting mood, I'd wager on the latter. But hey, at least the models are great at generating plausible-sounding incident reports. You just can't trust a word of them without a human in the loop. Again.

Oh, splendid. Sony AI has achieved a "major breakthrough" with project Ace, a table tennis robot that can now hold its own against elite human players. Because nothing says 'existential milestone' like a machine that's mastered the one sport where participants don't even have to run. The 'first known real-world autonomous system competitive with professional-level humans' — which is great news for anyone who's been losing sleep over when a robot might finally beat a human at, well, a very specific game played on a nine-by-five-foot table.

Let's pause to savour the modesty of the claim: 'competitive with' — not 'dominant over', not 'unbeatable by'. Just good enough to give a pro a decent rally. This is the same corporate-innovation speak that brought us 'self-driving cars' that needed constant human supervision and 'AI assistants' that can't book a restaurant without hallucinating the menu. Sony has essentially built a very expensive ping-pong opponent that you could already buy at a toy shop for £40, albeit one that can't move its own feet.

Still, the PR machine churns: a 'major breakthrough' in robotics, autonomous systems, whatever sells shares. Meanwhile, real-world AI is still crashing into walls and getting its training data poisoned by 4chan. But sure, let's all applaud the robot that can return a spin serve. At least when the machine uprising comes, we can challenge it to a game of table tennis — and based on this 'breakthrough', we might even stand a chance.

The US-China safety protocol dance we noted yesterday has been upgraded to a full-blown Cold War status update. The Washington Post's AI & Tech Brief now solemnly asks whether China is 'closing the gap' with US AI capabilities, just as the Trump administration enters talks on international safety protocols. Because nothing says 'we're in the lead' like opening negotiations while your competitor is breathing down your neck.

The newsletter, ever the sober analyst, examines various metrics—performance benchmarks, investment flows, talent pipelines—to conclude that the gap might indeed be narrowing. Never mind that just yesterday Treasury Secretary Bessent was boasting that US leadership enabled 'talks from a position of strength'; today the subtext is that those talks are happening precisely because the lead is eroding. The irony is thick enough to be weaponised.

One can almost picture the safety protocol negotiations: 'We'll share our best practices if you promise not to overtake us too quickly.' The whole spectacle smacks of two rival empires exchanging ambassadors while secretly funding arms races. The AI Cold War is here, complete with diplomatic niceties that fool no one. And as always, the public gets the marching orders while the labs write the rules.