The News,
Decoded.

So Senator Bernie Sanders, fresh from his many crusades against the billionaire class, has decided that the way to handle artificial intelligence is for the federal government to simply own a chunk of it. His plan: a one-time 50% tax on AI company shares, effectively transferring ownership to the public. Because nothing signals 'responsible innovation' quite like a confiscatory levy that would scare off every venture capitalist from Boston to Beijing. This follows the delightful dance between the Trump administration and OpenAI over a possible government stake — but whereas Trump's people were at least trying to negotiate a price, Sanders has dispensed with the market entirely. Why haggle when you can just take?

The WSJ opinion piece — titled with admirable candour 'The Road to AI State Socialism' — doesn't bother to hide where this leads. One imagines Sam Altman waking up to find his shareholder registry suddenly includes the United States Treasury as a majority owner. The 50% tax would presumably be applied to shares held by founders, investors, and employees alike, which ought to make talent retention a joy. 'Congratulations on your Series B — you now owe Uncle Sam half your equity.' And think of the message to the next generation of AI startups: incorporate in Delaware, get nationalised in Washington.

But let's give credit where it's due: at least the proposal is honest about its intent. While the White House talks about 'responsible AI' and 'national security', Sanders has simply cut to the chase. Nationalise the companies, seize the upside, and let the Pentagon sort out the rest. As far as regulatory gaps go, this is less a gap and more a bloody great chasm filled with tax forms and committees. The free market's response will surely be robust — probably through the medium of capital flight and a booming AI sector in Singapore. Cheers to progress, comrade.

So the White House has discovered that AI for national security needs to go faster — and has issued a statement promising to accelerate it, while also promising not to use it for "unlawful surveillance." On the very same day, Democrats introduced a 'responsible' defense AI bill we covered earlier — a framework so toothless it could double as a gum shield. One imagines the legal team working overtime to define "unlawful" in time for the machines to decide it themselves, given that the directive offers no enforcement mechanism and no independent oversight board.

This acceleration comes hot on the heels of the NSA quietly deploying Anthropic's Mythos for cyber operations — a model whose safety features are presumably being "gated" just enough to pass the sniff test. Around half-a-dozen Anthropic engineers are now at Fort Meade figuring out how to turn a "responsible" AI into a weapon. But hey, the White House says speed is key, so let's not let pesky questions about oversight slow us down. The fact sheet — yes, a fact sheet, not an executive order — provides no specifics: no budget, no timeline, no new oversight body. Just vibes and velocity.

And the "unlawful surveillance" promise? That's the most British of all signals — like a pub landlord saying "no trouble please" while the bouncers crack their knuckles. The directive includes no enforcement mechanism, no independent oversight board, and no timeline for publishing those "ethical safeguards." But don't worry — the White House stressed it. That's the important part. Between the NSA's secret integration of Mythos and the White House's vague acceleration pledge, the message is clear: we're going fast, and we'll figure out the rules later. The Dems' bill, meanwhile, calls for a "framework" but no binding restrictions. What could possibly go wrong? Probably nothing. Right?

So the Trump administration, which has spent years threatening to break up big tech and railing against government overreach, is now in talks to buy a piece of OpenAI. Sam Altman and the White House have been mumbling about a possible government stake for more than a year — a timeline that suggests either breathtaking indecision or a slow-brewing plan to nationalise the most hyped AI lab on the planet. The cognitive dissonance is enough to make a Daily Mail columnist weep into his flat white.

Let's game this out. A government stake means a board seat — or several — for Trump appointees. The same crowd that thinks climate change is a Chinese hoax will now help steer superintelligence. Possibly they'll demand it be trained exclusively on Truth Social posts and Fox News transcripts. And who pays? Taxpayers, naturally, whether the stake yields returns or not. The logic, as ever, is 'national security,' which in practice means 'we want to know what it's doing before it replaces us.'

This is how we get a state-run AGI that's as agile as the DMV. Altman gets a sugar daddy with infinite pockets; the White House gets a trophy asset. The rest of us get to watch a slow-motion collision between Silicon Valley's libertarian fantasies and Washington's bureaucratic grip. Cheers — that'll be another year of talks, I'm sure.

Oh good, the Democratic lawmakers have finally done it — they've introduced the 'Responsible Artificial Intelligence in Defense Act' on Monday, a bill that promises to establish a 'framework' for Pentagon acquisition and use of AI-enabled autonomous weapon systems. Because nothing says 'serious regulation' like a framework that hasn't passed either chamber yet and can be quietly amended into irrelevance by defence contractors' lobbyists. The manual override clause is particularly delicious — as if the machine that decides to fire a missile at a wedding party is going to politely wait for a human to press the 'nope' button.

The bill's core conceit is that human oversight and manual override will somehow tame the very systems designed to operate at machine speed. It's the legislative equivalent of putting a 'do not use in case of war' sticker on a tactical nuke. And let's be honest: the Pentagon has been acquiring autonomous systems for years — the 'framework' is essentially trying to bolt a horse-and-cart brake onto a hypersonic missile. The fact that this is being touted as a breakthrough tells you everything about where we are in 2026.

What's truly impressive is the sheer optimism that a single bill, introduced by one party, can close the gap between AI capability and regulation. Meanwhile, the systems are already being tested, deployed, and — as we've seen with the NSA's Mythos deployment last week — actively integrated into military operations. But sure, let's have a debate about a manual override. That'll work wonderfully.

Ah, the New York Department of Financial Services has issued an advisory warning that frontier AI models 'may amplify the potency, scale, and speed' of threat actors. Because nothing says 'decisive action' like a politely worded suggestion that something bad might happen. Insurers, who have spent years treating AI governance as a box-ticking exercise, are now expected to 'enhance their governance frameworks'—presumably by adding a few more bullet points to the same dusty risk register.

Let's not pretend this is new. The advisory is essentially a regulatory memo telling insurers that the sky might be falling, but please don't stop paying premiums. The specific threat vector? Frontier models helping threat actors 'identify vulnerabilities and exploits' faster than ever. As if insurers weren't already behind on patching their own systems. The DFS might as well have said: 'Hey, remember that fire? It's getting bigger. Carry on.'

The real joke is that this advisory carries no new enforcement teeth. It's a finger-wag, not a fist. Meanwhile, frontier labs keep releasing models with capabilities their own safety teams can't fully characterise. But sure, let's ask insurers to be more 'proactive'—because that's worked so well for data privacy, cyber resilience, and every other regulatory warning of the last decade.

Mustafa Suleyman, the man who once co-founded DeepMind before jumping ship to Microsoft, has announced that his division is now "set free" from OpenAI to pursue superintelligence. Because nothing says freedom like a contractual renegotiation with your largest investment partner, who you've already pumped billions into. The change happened about six months ago, which Suleyman helpfully disclosed in a recent interview—presumably after the lawyers had finished polishing the non-disparagement clauses.

Now Suleyman wants us to believe that was all just a prelude to the real show: Microsoft's own superintelligence push, powered by "custom silicon" and its own researchers. Because clearly, the best way to build something smarter than anything that exists is to clone the team that already did it, then declare independence like a teenager who still wants the allowance.

The timing is impeccable, coming just months after OpenAI's near-collapse and boardroom drama. Suleyman's cheerful announcement that Microsoft is "set free" to pursue AGI sounds less like a strategic pivot and more like a hostage negotiation where the hostage has Stockholm syndrome. Meanwhile, regulators are still trying to figure out whether Microsoft's investment in OpenAI constitutes a merger. Spoiler: it does, but nobody's got the stomach to enforce that. So yes, Microsoft is free—free to race toward the same cliff as everyone else, just with its own branded silicon. Cheers.

So SK Telecom has joined Anthropic's Project Glasswing, gaining access to the same Claude Mythos Preview that the NSA is reportedly readying for cyber operations. Because nothing says 'responsible AI deployment' like having your safety-gated model piloted by both a signals intelligence agency and a South Korean telecom giant. Two very different customers, one delightful implication: the gating works only until someone decides the gate should open for them. The release notes must be a hoot: 'Safety measure: limit usage to NSA and SKT only — two parties with impeccable track records of not abusing surveillance capabilities.'

The model is meant to proactively identify vulnerabilities in SKT's core infrastructure — a mobile network that handles data for tens of millions of subscribers. Let's hope it's better at that than it is at distinguishing between 'adversarial nation-state traffic' and 'Seoul commuter streaming video.' Anthropic's PR wraps this in language about 'collaborative partnership' and 'secure access,' but we all know this is an audition: prove Mythos can protect a telco's crown jewels, and next it'll be banks, power grids, your smart fridge. The 'gated' descriptor starts to lose meaning when every major institution wants a key.

The really delicious part? Project Glasswing is supposedly about 'gated' deployment — access controlled by Anthropic, safety first, all that. Yet here we are, just weeks after the NSA story, with a second major, opaque deployment of the same model, this time to a foreign telecom. The gate seems to have a revolving door, and the hinges are greased with 'strategic partnership' cash. What could possibly go wrong when a frontier cybersecurity model gets its hands on real-time telecom infrastructure telemetry? Oh right, everything — starting with the small matter of zero-day exploits being found and potentially exploited by the model's operators before anyone else even knows they exist.

So Anthropic — the lab that solemnly calls for pause mechanisms and signs letters against bioweapons — is now helping the NSA weaponise its cybersecurity model. Around half-a-dozen engineers have been deployed to Fort Meade to ready Mythos for cyber operations. "Frontier cybersecurity AI," they call it. Sure. And the CIA's "preventative detention" programme was just administrative housekeeping.

The Financial Times report, picked up by TechCrunch, notes this is about Mythos — a model Anthropic has been pitching as the responsible answer to escalating cyber threats. No doubt it is. The same kind of responsible that saw Cambridge Analytica use Facebook data to "help" elections. When the world's largest intelligence agency embeds engineers from a safety-focused lab, you don't get oversight — you get a press release about "partnership in national security."

Let's be clear: Mythos might be brilliant at detecting vulnerabilities. It might even be used defensively, for now. But once the NSA has its hooks in a frontier model, the line between defence and offence gets very blurry, very fast. And Anthropic, the company that wanted an industry-wide kill switch, just handed the NSA the keys to the engine. Pass the sick bag, Alice.

Microbiologist Elisabeth Bik, the Dr. Watson of scientific fraud detection, has raised the alarm: generative AI is now fabricating research data on an industrial scale. Her warning, reported in Research Professional News, confirms the dirty secret that academic publishing has been dreading — that the machines have learned to lie convincingly. The 'industrial scale' phrase is a beautiful euphemism for 'we are completely fucked', and Bik knows her stuff. She has been chasing manipulated images for years; now she faces a synthetic firehose.

Consider the implications. Every day, hundreds of papers flood into journals, many already borderline in quality. Now add a bottomless well of AI-generated tables, charts, and statistical tests that look perfect to the tired eyes of overworked reviewers. The labs building the next trillion-parameter model will happily train on this enriched corpus, baking the hallucinations into future systems. It is a feedback loop of bullshit that would make a data scientist weep — if they weren't busy generating more of it. And the cycle accelerates as more papers get accepted, more datasets get contaminated, more AI models learn from the very garbage they produce.

And what of the brave regulators who will one day save us? They are still debating whether AI poses an existential risk, while the scientific record itself is being quietly poisoned. Bik's sleuthing is heroic, but she is one person against a tsunami of synthetic data. The joke, as ever, is that we spent billions building a digital forger and then complained when it started forging. Cheers, progress. At least the PhD thesis mills and predatory journals will have fresh raw material to keep the academic racket running smoothly.

Anthropic, the lab that last week solemnly called for an industry-wide pause mechanism, now proudly announces that Claude writes over 80% of its production code. The irony is thick: the very code running the models they want to pause is mostly authored by the model itself. They even published a guide on 'how your enterprise can keep up,' encouraging others to hand their production keys to an AI they've flagged as potentially dangerous. Anthropic's own engineers now spend their days reviewing Claude's output — a role that suggests they trust the quality but not enough to walk away entirely.

The 80% figure is an impressive technical feat. In May, Claude generated and merged code into Anthropic's codebase at a pace humans can't match. Anthropic calls it a productivity breakthrough, but it makes their earlier pause proposal look like a cynical PR exercise. The same engineers who trained Claude are now rubber-stamping its pull requests, effectively training their own replacement and calling it progress.

What of the broader workforce? The enterprise advice boils down to: let Claude write your code too — what could possibly go wrong? The subtext is clear: if we do it, it's safe. Yet they simultaneously argue for coordinated brakes. It's a glorious contradiction: safety advocate on Monday, automation evangelist on Tuesday. The human loop is tightening by the day. Time to pour yourself a pint and try not to think about it.

Marvelous. Generalist AI Inc. — a name so blandly ambitious it could be the title of a PowerPoint slide — has bagged $400 million from investors who apparently believe that one model to rule all robots is just around the corner. Total funding now stands at over half a billion dollars, which is roughly the GDP of a small country that could actually use functioning robotics. But never mind that: the press release says this cash will "scale its general-purpose AI models for robotics," which is startup-speak for "we haven't figured out the hard bits yet, but here's a big cheque."

The general-purpose promise is the holy grail of robotics: a single AI that can drive a forklift, flip a burger, and perform surgery without requiring three years of retraining. In practice, robotics is a graveyard of overpromises, littered with the carcasses of startups that thought they'd cracked the code. Generalist’s believers are betting their $400 million that this time is different — because the fifth time's the charm, obviously.

And what of the robots themselves? But until a robot can reliably pick up a toaster without destroying it, the word "general" should probably come with an asterisk. Still, $400 million buys a lot of press coverage, and the CEO gets to say "we're solving general intelligence for robots" with a straight face. Which, in this industry, is essentially the product.

Oh, marvellous — the two high priests of the AI church have found common ground, and it only took the spectre of engineered pandemics to do it. Sam Altman and Dario Amodei signed an open letter asking policymakers to regulate synthetic nucleic acid ordering, because nothing says 'alignment' like begging for laws after your labs have already unleashed the tech. The letter is a masterpiece of timing: Altman is still smarting from the GPT-5 safety debacle, and Amodei — who spent years wagging his finger at everyone — now needs a government leash to look serious.

The real kicker? This isn't about preventing bioweapons — it's about appearing to. The actual mechanism — gatekeeping DNA synthesis — is about as enforceable as a library's 'please be quiet' sign. If a bad actor wants to cook up a pathogen, they'll either order from a Chinese supplier or synthesize it in a garage. But sure, a letter will stop them. Meanwhile, Anthropic's brilliant coordination plan from last week is already gathering dust, and Altman is back to talking about AGI by 2027.

So we get a press release dressed as statesmanship. Two CEOs, identical twins in hubris, finally agreeing on something that costs them nothing. They'll write letters, make speeches, and go back to building ever-faster reasoning models that can, oh by the way, design novel toxins. The bioweapons letter is a lovely distraction — a bit of regulatory theatre to keep the concerned citizens mollified while the real work of maximizing compute continues. Cheers, lads. Try not to sneeze.

The House subcommittee has been officially warned that AI is dangerous and might need regulating — a conclusion that required a whole hearing, multiple witnesses, and presumably several cups of bad congressional coffee. The kicker, as reported, is that rules 'might not stop foreign adversaries from copying dangerous AI models and removing guardrails.' So we're regulating the honest labs while adversaries just download the weights and strip the safety tags. It's like securing your front door with a paper sign saying 'Please don't rob me' while leaving the back open to anyone with a USB stick.

The witnesses did their duty: they described the threat landscape, mentioned cyberattacks and disinformation, and politely noted that open-source models are basically untraceable. No one proposed a technical solution because there isn't one — at least not one that doesn't involve either global surveillance or unplugging the internet. So the hearing will generate a transcript, a few think-pieces, and precisely zero change. The gap between the warning and the response is so wide you could drive an entire adversarial nation through it.

And what of the 'cautious policy response' the title demands? Cautious means slow, and slow means too late. By the time Congress drafts a bill, the foreign adversaries will have already copied, stripped, and deployed the next generation of models. The subcommittee might as well have held a séance to summon a regulatory framework from beyond the political grave. But hey, at least we had a hearing. That counts for something in Washington — about as much as a guardrail on an AI model that's already been forked a thousand times.

Ah, the U.S.-China AI race just got a new player: Yao Shunyu, a former OpenAI researcher who has decamped to Tencent as chief AI scientist with the modest ambition of building AGI. Because nothing says 'responsible development' like moving the goal from 'alignment' to 'global supremacy' via talent poaching. CNBC reports that China is now actively luring ex-OpenAI staff, presumably because if you can't beat 'em, hire 'em. The irony is exquisite: the same lab that warned about catastrophic risks is now a recruiting ground for a state-backed rival.

Yao's declared aim to achieve AGI at Tencent is a direct challenge to OpenAI's mission, but let's not pretend this is about scientific curiosity. This is about who gets to define the rules of the game. China's AI firms are reportedly offering salaries that would make Silicon Valley blush, and the exodus of talent is accelerating. The US spent years worrying about China catching up on hardware; turns out the soft underbelly is the researchers themselves. And what's the American response? A few grumpy op-eds and a half-hearted executive order that'll be forgotten by lunch.

So here we are: the world's two biggest economies locked in a talent war for the people most likely to build something we can't control. And the chap who left OpenAI for Tencent says he'll 'open-source AGI if it's safe' — which is about as reassuring as a parachute packed by a three-year-old. The irony is that the US created the environment where AI talent flourishes, then promptly forgot to make them stay. China just needed to offer a bigger pot of gold and the promise of fewer ethics committees. We're in for a glorious race to the bottom.

Anthropic, the lab that wants you to think it's the responsible one, has proposed a coordination plan for AI developers to pause if risks rise. Because nothing says 'we're serious about safety' like a voluntary, trust-based mechanism in a race where the gold medal is AGI. The proposal, reported by Reuters, suggests labs should agree on thresholds to slow or halt development if systems begin 'self-improving faster than society can manage risks' — a beautifully vague phrase that assumes society is currently managing risks at all. It's like asking a pack of greyhounds to voluntarily stop chasing the mechanical hare because it might be too fast.

The irony is thick enough to bottle: Anthropic itself is a frontier lab racing toward capabilities that could trigger the very scenario it warns about. The 'coordinated plan' is basically a polite suggestion in a ruthless competitive landscape where any pause risks ceding the lead to DeepMind, OpenAI, or a state-backed player who didn't get the memo. And let's be honest, 'coordinated' in this industry usually means 'we'll talk about it over coffee until someone releases a model that makes the discussion moot.'

The proposal is probably sincere — Anthropic's safety-focused brand depends on it — but it's a bit like asking the arsonists to install fire extinguishers while they're still dousing everything in petrol. The only coordinated halt we'll see is when something actually goes wrong, at which point this plan will be retroactively hailed as prescient and ignored. Cheers to proactive governance, lads.

Ah, January 2026: the month Teradata's 5,100 employees learned their annual salary raise was being sacrificed to the AI gods. The message is clear: you, the worker, are a cost centre; the chatbot, a divine investment. No raises for you, but rest assured, the model will be well-fed. Because nothing says 'we value our people' like telling them their compensation is now a trade-off against a probabilistic autocomplete engine. The gall of framing this as a reallocation rather than a wage freeze is the kind of creative accounting that would make an Enron CFO blush.

Teradata's decision is a microcosm of a broader trend: companies openly treating human salaries as a variable cost to be slashed in favour of AI subscriptions. The message to employees is not just 'you're not getting a raise' but 'your career progression is now optional.' At 5,100 people, that's a lot of disillusioned talent wondering if their next performance review will be conducted by a large language model. And the fun part? There's no guarantee this AI investment will yield a single dollar of profit. It's an act of faith, paid for with your pay rise.

But hey, at least the CEO can tell shareholders they're 'investing aggressively in AI' while the workforce stagnates. The workers, meanwhile, get the honour of funding the very technology that may replace them. It's the corporate equivalent of asking the condemned man to pay for his own rope. Teradata is not alone, but they've set a lovely precedent: 'Sorry, no raise — we're buying a shiny new AI.' Cheers to progress. Same as it ever was.

Ah, the factory floor has discovered the ultimate paradox: efficiency so dizzying it gives you time to breathe, and that’s precisely what’s terrifying. A worker quoted in 2026 says they “have time to breathe again” thanks to AR/AI tools that have cut order processing by a crisp 4x in some pilots. Because nothing says job security like demonstrating you can complete your entire day’s work in a quarter of the time. The next logical step, of course, is for management to conclude that three of every four workers are now surplus to requirements. The anxiety isn’t a bug; it’s the feature they haven’t rolled out yet.

The article doesn’t name the factory or the tool vendor — probably because the story is depressingly universal. AR headsets and AI copilots are being deployed everywhere to “empower the workforce” while quietly compiling a spreadsheet of who can be safely let go. The 4x figure is proudly cited, as if speed alone were an unqualified good, without a nod to the fact that when every task takes a quarter of the time, the headcount can follow suit.

So here we are in 2026, with the blessed gift of downtime becoming a source of dread. The same forces that gave us the gig economy and the 24/7 inbox have now weaponised spare minutes. The worker who finally breathes is the one who’ll be first out the door when the ROI calculations land. Next year’s pilot will likely optimise breathing itself — we’ll hear “I have time to blink again” before the lid is sealed. Cheers to progress — may your next deep breath be your last at this company.

Ah, the American Public Transportation Association has convened its 2026 Workforce Summit to ask the truly pressing question: 'Is your workforce AI ready?' Because nothing says 'prepared for the autonomous future' like a summit that couldn't even spring for a snappier tagline. The focus, predictably, is on AI and autonomous vehicles in public transport — because what could possibly go wrong when we hand over the control of city buses to a model that might interpret 'stop sign' as 'suggestion'? One imagines the keynote speaker reassuring the audience that retraining programmes will definitely cover the thousands of drivers whose jobs will evaporate.

Of course, the APTA's timing is impeccable. The summit probably included a workshop on 'Upskilling for the Autonomous Era' — because nothing says 'we value our workers' like a PowerPoint presentation on how to become a 'fleet monitor' earning half your previous wage. The implicit message is clear: your job will be automated, but you can still feel good about it if you attend enough sessions.

Let's be honest — this summit is a masterclass in performative concern. The APTA knows full well that autonomous vehicles are coming, but rather than confront the actual disruption, they've organised a talking shop. The outcome will likely be a white paper, a few 'best practice' guidelines, and a collective sigh of relief that no actual decisions were made. In the meantime, the workforce is left with the warm fuzzy feeling that someone, somewhere, is asking if they're ready. Spoiler: they're not, and a summit won't fix it.

Ah, the tech sector's relentless romance with efficiency continues. US technology companies announced 38,242 planned job cuts in May — the highest since August 2024, according to Bloomberg. Because nothing says 'confident about the future' quite like showing 38,242 people the door, all in the name of ramping up AI spending. The logic is impeccable: replace expensive humans with expensive algorithms that may one day replace even more humans. It's a self-fulfilling prophecy wrapped in a press release.

And what a charmingly circular economy this is. Companies cut staff to plow cash into AI, which then gets used to automate the very roles they just cut. The 38,242 figure is just a down payment on the next round of 'efficiency gains.' Bloomberg notes this is the most since August 2024, which tells you the pattern is not a blip: it's a feature. Every spike in AI investment is met with a spike in pink slips, and somehow the industry keeps a straight face.

But fear not — those displaced workers can always retrain for the booming field of 'AI ethics' or 'prompt engineering,' right? The gallows humour writes itself. Meanwhile, the executives who approved the cuts will be on conference stages next week talking about 'human-centric innovation' with perfect sincerity. If there's a more reliable indicator of impending social friction than record job cuts paired with record AI spend, we'd love to see it — just send it to our LinkedIn inbox, which is about to be very, very quiet.

JFrog's research proclaims that 68% of Australian organisations self-host AI models — the highest rate among countries surveyed. Because nothing says 'secure, sovereign AI' quite like handing your most sensitive data to a model running on a server in someone's broom closet. This isn't just a statistic; it's a monument to the charmingly naive idea that keeping the model on-premises somehow makes it safer, when in reality you've merely relocated the vulnerability from a cloud provider's server to your own increasingly sweaty palms.

Of course, the survey doesn't reveal what these organisations are actually doing with all those self‑hosted models. Are they generating personalised spam, hallucinating compliance documents, or merely hoarding GPUs for winter? JFrog, a company whose entire business model is finding and fixing security holes in software, is naturally delighted that so many Australians are now running AI that needs securing. The subtext is unmistakable: 'Congratulations on your new self‑hosted LLM — mind if we run a vulnerability scan?'

This is the Australian AI sovereignty dream made real: total control, total responsibility, total unpredictability. While the rest of the world bickers over watermarking and model licenses, the Land Down Under has quietly become a vast petri dish for unsupervised AI. It's not paranoia if the models really are hallucinating your bank statements. She'll be self-hosted, mate.