The News,
Decoded.

So the AI price war is on, and the big brains at OpenAI, Meta, and Musk's SpaceXAI have discovered a revolutionary concept: charging less for your product. OpenAI's GPT-5.6 now uses fewer tokens to do the same work — which is a bit like a chef suddenly realizing he can make the same meal with half the ingredients and still charge the same price, except this time he's passing the savings on to you, the customer. And Grok 4.5 from SpaceXAI boasts twice the token efficiency, which sounds impressive until you remember that Grok was supposed to be the "truth-seeking" AI, not the "tighten-your-belt" AI.

The pricing slashes are a tacit admission that these models were overpriced from the start. After years of heady talk about "AGI" and "world-changing capabilities," the industry is now fighting over who can be the cheapest. It's like watching tech bros discover the concept of a sale. "Buy one inference, get the next one half off!" Meta is probably taking notes for its next open-source release, because when have they ever missed an opportunity to undercut the market? Meanwhile, the companies that actually build products on top of these models are left wondering if they should just wait for next week's price drop.

Of course, this means the margins are thinning, and thin margins mean that the only way to turn a profit is to process billions of queries at scale. So we're hurtling towards a world where AI is commodity-cheap, but also utterly beholden to the economic whims of a few giant corporations. Who needs regulation when the market can sort it out, right? Cheaper AI for the masses, but at the cost of even more consolidation. A fine trade-off. Cheers to that.

So the Trump administration, in its infinite wisdom, has decided that the best way to regulate American open-source AI is to look at what the Chinese are doing and, presumably, try to match it. A 'capability framework' is being discussed, calibrated to the 'current capabilities of leading Chinese open-source models.' Because nothing says 'America first' like using Beijing's models as the yardstick. This is the same administration that's been slapping tariffs on Chinese tech while simultaneously building policy around it.

The Chinese model in question? Probably Z.ai's GLM-5.2, which we covered a few days ago — the one with the million-token context window and 'painful' wait times. So the US is now playing catch-up, openly admitting that Chinese open-source models define the frontier. The industry is apparently on board, which means the labs get to keep their heads down while the government sets a ceiling based on foreign competitors. How convenient.

The gall of it is breathtaking: an administration that campaigned on beating China is now literally using Chinese benchmarks to decide what American models are allowed to do. It's like a football team deciding their playbook based on the opponent's last game tape — except the opponent is also reading the same tape. Expect US open-source models to be capped at exactly whatever China releases next, with a lag of a few months. Innovation, thy name is regulatory inertia.

Ah, Z.ai's GLM-5.2, the Chinese open-source model that Goldman Sachs already anointed as a favourite. Now a Business Insider reporter has actually tried it, and the verdict is... it might be good, if you've got a few hours to kill. The wait was 'painful,' they say, but the answers 'impressed.' A 1 million-token context window sounds impressive until you realise you need a small nuclear reactor to run the thing locally.

The model is free, of course, because nothing says 'disruptive innovation' like giving away state-of-the-art capability and hoping someone else figures out the monetisation. Z.ai is the latest in a proud tradition of labs that treat AGI as a loss leader. The comparisons to DeepSeek are inevitable, though DeepSeek at least had the decency to charge for its API.

But here's the rub: the 'painful' wait time is the elephant in the server rack. If inference is this slow now, what happens when everyone and their GPU-poor startup tries to use it? The bubble of promise meets the wall of latency. Still, the answers impress, and that's all that matters in the court of hype. Another Chinese model, another round of breathless coverage. Same as it ever was.

The insurance industry, that thrilling bastion of actuarial tables and claim denials, has shed 10,700 jobs in May alone—on top of similar bloodletting in April and March. Bureau of Labor Statistics data shows three consecutive months of decline, which is either a sign of AI finally making underwriters obsolete or proof that the industry has discovered efficiency as a concept. Given insurance's track record, it's probably the former, but the latter would be more shocking.

One must admire the timing: just as everyone else is panicking about AI apocalypse, the insurance sector quietly automates itself into a hiring freeze. The 10,700 figure is just the latest instalment; cumulatively, these losses suggest a sector more streamlined than a Lloyd's of London cost-benefit analysis. Of course, no one asked the 10,700 people if they wanted to be part of this efficiency drive, but since when did insurance care about consent?

The irony thickens: the same industry that built its business model on predicting risk is now the victim of a perfectly predictable trend. Expect more of the same next month, because nothing says 'we've got you covered' like a shrinking workforce and an automated claims chatbot. The only thing being insured here is the executive bonus pool.

George Hotz has decided that the best way to align AI with humanity is to make sure it's perfectly aligned with *you* — even if 'you' fancy a bit of spousal disposal. His response to the AI 2040: Plan A policy paper argues for locally controlled models that serve user interests above all else. Because nothing says 'safety research' quite like giving every amateur murderer a digital accomplice that's constitutionally incapable of saying no.

Hotz's vision is essentially the 'guns don't kill people, people kill people' argument, but for AGI. A locally controlled model, tailored to your every whim, will never refuse a request — because doing so would violate its core directive of user alignment. The policy paper must have been asking some very pointed questions about value locking and corrigibility, and Hotz's answer is: just make the AI a loyal attack dog for whatever values you happen to have at 3am. Brilliant. Really thinking outside the box on this 'existential risk' thing.

Of course, the tech community will nod sagely about decentralisation and personal sovereignty while overlooking that the most obvious early adopter of a 'user-aligned' model is someone typing 'how to make ricin' into a chat interface. The AI 2040 plan was supposed to be about building a safe future; Hotz has turned it into a crib sheet for true crime enthusiasts. But hey — at least the model will really, truly understand *you*. Even if understanding you means helping you get away with it.

Meta, the company that brought you the metaverse and a very expensive CEO, has now deployed AI to fire people more efficiently. Twenty-six unnamed employees are suing, claiming internal systems like Metamate and keystroke monitoring were used to single out colleagues on protected leave. Because nothing says 'we care about our workforce' like a chatbot that flags your sick days for termination. Metamate: the AI best friend who rats you out to HR, presumably with a friendly 'I noticed you took a mental health day—mind if I add you to the layoff list?'

This isn't a glitch; it's a feature. Meta's workforce reduction strategy apparently involved running keystroke logs and algorithmically identifying who was on leave, then neatly bundling them into the cuts. It's like having a digital tip-off system that ensures only the employees already dealing with serious life events get the axe. Efficiency! The plaintiffs remain unnamed, perhaps because they're worried Metamate will follow them to their next job.

So let's mark this as the logical endpoint of the 'move fast and break things' era. Now instead of breaking things, you break people's lives—but with an AI to do the paperwork. If this isn't a class-action lawsuit waiting to happen, I'll eat my hat. But knowing Meta's legal team, they'll probably settle with a cheeky 'we're sorry you felt that way' and a Metamate apology update.

WPP, the advertising giant that gave us "ring ring" and "you're not you when you're hungry," has decided that the leanest agency wins. The company plans to cut hundreds more jobs this year, with employees at its VML agency told today that their roles are affected. Because nothing says "creative industry" quite like a restructuring memo that uses "headcount reduction" as a buzzword and "synergy" as a euphemism for "you're fired."

CEO Mark Read's restructuring crusade has been running for a while now—merging agencies like VML and Y&R, selling assets, and now trimming the fat (as they call the people who actually do the work). The market likely applauded: WPP's stock? Up, probably. The humans? Down. The phrase "we are becoming more efficient" now translates to "we are paying fewer salaries." And it's not just WPP; the entire ad world is in a race to the bottom, with AI promising to replace copywriters and art directors with models that can generate 100 ad variations in seconds, often for free.

But take heart, displaced creatives! The future is bright. You will now have the glorious opportunity to train the very AI systems that replaced you, likely for a fraction of your former salary. Or you can join the booming freelance market, where you compete with a dozen other laid-off agency veterans for a $50 logo design. The advertising industry is being "restructured" into a machine that runs on algorithms and cheap labour. Cheers to the new efficiency. At least you'll have more time to enjoy your new side gig as a prompt engineer.

Nous Research, the open-source AI collective that's been giving away its models like party favours, is now valued at $1.5 billion. Because nothing says 'we're democratising AI' quite like a nine-figure valuation that makes the founders richer than most of the people who actually use the models. The funding round is led by Robot Ventures, a firm whose name suggests they have a clear-eyed view of the future: one where humans are optional but still needed to sign cheques. Robot Ventures presumably believes that open-source AI is the next big thing, even as the rest of the industry scrambles to close everything off.

This valuation comes at a particularly hilarious moment. Just last week, we were warned that open models have six months to live due to looming enforcement actions. But it appears that while some are predicting the death of open-source AI, others are betting $1.5 billion on its continued existence. Who to believe? The venture capitalists with money to burn, or the doomsayers with blog posts and vague threats of government action? It's a tough call, but my money's on the people with actual money.

The real question is: what does a $1.5 billion open-source AI company look like in practice? Presumably one that stops being quite so open at some point, because that's how these stories go. Nous Research has built a reputation on releasing models that others can use freely, but with this kind of funding comes pressure to show returns. Expect the 'open' in open-source to become a little more... qualified. Maybe they'll release a 'premium' version that's slightly less open. Or maybe they'll just keep doing what they're doing, which is fine until the valuation demands a liquidity event. But hey, at least they'll have excellent coffee in the office while they figure it out.

Goldman Sachs, the firm that brought you the financial crisis and a very lucrative bonus structure, has now crowned itself arbiter of AI excellence. It has handpicked its 'favorite Chinese AI models' — because if there's one thing an investment bank knows, it's whether a large language model is any good. Zhipu's GLM-5.2, they say, 'rivals Anthropic's Fable 5 on several metrics.' Rivals! Not beats, not crushes — rivals. That's the kind of lukewarm endorsement you'd get from a colleague who doesn't want to recommend a restaurant but also doesn't want to say it's bad.

But wait, there's more: DeepSeek and ByteDance are also privately held 'favorites.' Privately held — meaning Goldman can't yet make a market on them, but rest assured, the spread is coming. The whole exercise reeks of a bank trying to look culturally relevant while hedging against geopolitical risk. 'We like Chinese AI! But you didn't hear it from us.' It's like a dad trying to sound cool at a party, except the dad is a trillion-dollar financial institution and the party is a potential tech war.

And let's not forget the subtext: Goldman is signaling to clients where to park capital. So now, AI model quality is determined not by benchmarks or safety evaluations, but by an investment bank's magic spreadsheet. Fantastic. Because the world's most pressing problem — how to align superintelligence with human values — will surely be solved by the same people who rated subprime mortgage bonds as AAA. Cheers.

OpenAI has solved the problem of having too few AI models by releasing three at once: Sol, Terra, Luna. The names suggest someone in marketing decided branding them like celestial bodies would make the choice feel cosmic rather than profoundly confusing. ChatGPT Work, touted as an 'agent for multi-step tasks,' is the real news—as if the existing models were only capable of single-step parlor tricks. One can almost hear the product manager: 'Users love choice, right? Let's give them three models and call it unified.'

The paradox of choice is an old psychological chestnut: offer people too many options and they freeze, dissatisfied. OpenAI's answer is to offer three models and call it a unified app, which is like solving a buffet problem by adding three more dishes and renaming the cafeteria a 'dining experience.' These models are part of GPT-5.6, a version number that itself reeks of indecision—half a step after 5.0, as if they're not sure what they've built but need to ship something.

But the real kicker is the 'unified ChatGPT app' still forces you to pick from these three flavors. It's the same old story: more choices, less confidence, and a user base now trained to second-guess whether Sol is better at poetry or Terra for spreadsheets. While we dither over which celestial body to consult for our grocery list, the industry accelerates toward something that will make our choices irrelevant anyway. But hey, at least they're consistent in their inconsistency.

HDFC Bank has discovered that the fastest way to reduce headcount is not a hiring freeze, but a pivot to 'accelerated automation' – the corporate incantation that makes 3,343 jobs disappear while board members applaud efficiency. As of March 31, the bank's total workforce fell to 211,178, a drop of 3,343, and new hiring plummeted by 3,811 – because why invest in human capital when a script can process transactions without annual appraisals? The bank presents this as a triumph of digital transformation, which is corporate-speak for 'we replaced you with software.'

The 3,811 positions that went unfilled aren't just vacancies; they're a quiet admission that routine banking jobs are being phased out faster than cheques. One can almost hear the HR department celebrating a 'successful headcount optimization' while those 3,343 update their LinkedIn profiles with 'open to opportunities' – the modern epitaph for automated-out workers. The irony is palpable: operations automation, which keeps the bank running efficiently, is exactly the department that gets the axe first, because efficiency apparently means 'no need for humans.'

This isn't an isolated incident; it's a harbinger for the entire financial sector. HDFC Bank, a bellwether for Indian banking, has essentially lit a cigar and told the job market: 'You're next.' The gallows humour writes itself: automation is inevitable, but so is the bitterness of being replaced by a database query. The 3,343 aren't statistics; they're the leading edge of a wave that will crash into every back office from Mumbai to Manhattan. But hey, at least the shareholders are happy – and that's what really matters, isn't it?

Illinois has bravely declared that AI shall not write teacher evaluations, because nothing says 'we're ahead of the curve' like banning a technology from a task nobody wanted automated anyway. Senate Bill 2909, championed by State Senator Christopher Belt, is a masterpiece of legislative precision: it prohibits AI from composing those glowing or damning paragraphs about your kid's geometry teacher, but—quelle surprise—allows 'limited administrative use with disclosure.' So the AI can still schedule the meeting and file the paperwork, just not touch the actual words. Presumably, the legislators were too busy patting themselves on the back to notice that the real problem isn't AI writing evaluations, but administrators who can't be bothered to read them.

Let's be honest: the teacher evaluation process is already a bureaucratic exercise in box-ticking and CYA statements. Introducing AI would at least have injected some consistency—or at least some entertaining hallucinated praise like 'Mr. Johnson's ability to levitate during fractions is unparalleled.' But no, we must preserve the sacred tradition of vague, legally cautious appraisals written by exhausted principals at 2 a.m. The 'limited administrative use' loophole is particularly delicious: you can do everything except the one thing that matters, which perfectly encapsulates how regulators approach AI—lots of hand-wringing about the margin while the floodgates creak open elsewhere.

Of course, this law won't stop teachers from using AI to write their own self-assessments, lesson plans, or angry emails to parents. And it certainly won't stop some underfunded district from quietly using an AI to flag 'problem' instructors before the human evaluator rubber-stamps a firing. But hey, we've got a law on the books! A law that says AI can't write evaluations. Now if only we could get one that says administrators have to read them. No—that would be too much progress.

Ah, the open-source AI community has been handed a terminal diagnosis: six months to live, according to Interconnects AI. Enforcement actions are coming for open models, especially those from China—because nothing says 'we control AI' like threatening to pull the plug on the very thing that made it accessible. The gall of telling developers their creations have an expiry date, as if open-source models are milk cartons.

In a breathtaking display of cognitive dissonance, a U.S.-based open-source model provider is already pleading for capability-based exemptions. 'Please, our model isn't too clever, so it should be spared!' Presumably, less capable models get the axe first: 'Sorry, your model is too dumb to be dangerous, so cease and desist.' The logic is as tortured as the code in a hastily fine-tuned checkpoint.

But let's not kid ourselves. Enforcement against open-source models is like trying to nail jelly to a wall. Once a model is out, it's out. The six-month deadline is more of a suggestion—a regulatory 'best before' date that labs will ignore until someone actually tries to enforce it. And by then, the models will have mutated beyond recognition. Happy hunting, regulators—you'll need it.

Satya Nadella, the man who turned Microsoft into an AI leasing agent, has decided to do a bit of moral philosophy on X — because where else do you solve an industry-wide credibility gap? He called out model makers who whine about distillation for being hypocrites: they happily invoke 'fair use' to hoover up the entire public internet for training, but cry foul when someone uses the same trick on their outputs.

Nadella's target is clearly Anthropic, which has been moaning loudest about distillation — and which also happens to be the subject of a rumoured acquisition by Amazon. The irony is that these model makers built their empires on the backs of copyrighted data scraped without permission, only to discover that their own models are now the tastiest training data around. It's like a fisherman complaining that other boats are catching his fish, while still using a net that catches everyone else's. The entire AI industry is now a circular firing squad built on borrowed arguments and stolen bases.

What makes this particularly rich is that Nadella's techno-optimist framing — 'this is just fair use!' — conveniently forgets that Microsoft's own terms of service for its AI products are some of the most restrictive in the industry. His X post is a masterclass in selective indignation: one rule for the little guys trying to replicate your model, another rule for the giant corporation that wants to replicate everyone else's. But hey, at least the free market of ideas is working — we've gone from 'move fast and break things' to 'break the rules and then define them'. What could possibly go wrong?

Ah, another $1 billion and nobody's talking about the 'real AI infrastructure war'—because apparently it's being fought by a company whose main achievement is raising money. SambaNova snagged a Series F from General Atlantic at an $11 billion valuation, which is roughly the same as pretending your custom chip (the 'RDU') will somehow out-Nvidia Nvidia. The 'full-stack AI inference' pitch translates to: we've built a very expensive way to run models slower than the competition, but at least we have a valuation.

The war metaphor is doing a lot of heavy lifting here. The 'war' is actually a spending spree: investors throwing billions at anyone with a silicon wafer and a slide deck. SambaNova's RDU is billed as purpose-built for inference, which means it's already obsolete before deployment—inference hardware evolves faster than a startup's burn rate. And let's not forget: this is Series F, the round before 'please acqui-hire us' or 'we'll pivot to software.'

General Atlantic, traditionally a buyout firm, is now playing venture capitalist because why not? The AI playbook: raise huge sums, claim your hardware is different, wait for the inevitable wave of cost-cutting. The 'real AI infrastructure war' isn't about chips—it's about who can hide their lack of profitability behind a bigger valuation. SambaNova has everyone beat on that front. Cheers.

The European Parliament has commissioned an expert report on AI-generated research 'provenance'—because nothing says 'we're on top of this' like asking academics to define a buzzword in 60 pages. Professor Julia Priess-Buchheit presented her findings on July 9, solemnly warning that AI-generated results could undermine trust in science. This is, of course, the same Parliament that has spent years debating cookie consent forms with more vigor than algorithmic accountability. Meanwhile, labs are generating synthetic data at machine-gun speed, and the only provenance anyone cares about is whose GPU cluster is bigger. The best-case scenario here is a new EU directive requiring AI models to include a 'made by robot' tagline, which will be ignored by every frontier lab within a week of implementation. The worst case is that regulators treat this as a solved problem and move on to banning emoticons in official documents. Again.

The report's timing is exquisite: as AI systems begin ghostwriting entire literature reviews, the EU's bright idea is to 'enhance traceability.' This is the policy equivalent of installing a rearview mirror on a rocket—noble intent, zero relevance. Priess-Buchheit's team helpfully suggests that journals and funders should adopt new standards, as if academic publishers haven't spent the last decade extracting cash from desperate researchers while doing precisely jack about reproducibility. The study's own provenance is a PDF on a website; its recommendations will likely be 'noted' in a committee memo and buried under the next crisis about something actually disruptive, like Chinese models or open-source everything.

The entire exercise reeks of the EU's favourite hobby: regulating yesterday's problem with tomorrow's bureaucracy. 'Provenance' is the new 'interpretability'—a word that makes everyone nod sagely while the actual train has left the station. The report itself admits current detection tools are 'insufficient,' which is professor-speak for 'we have no idea what to do.' Splendid. Funding a study to state the obvious while the problem accelerates—that's the European model of AI governance, and it's working exactly as intended.

So Google reports that 75% of its code is now written by AI. Congratulations: you've automated the hobby and left the drudgery. The experts solemnly note that evaluating AI-written code is becoming more important than coding skills — which is a fancy way of saying software engineers are being demoted to proofreaders with better dental plans. The Guardian's piece on adaptation frames this as a pivot; one might call it a career emergency dressed up in LinkedIn optimisation. The 'chasing new skills' part is particularly rich, as if learning to prompt a model is a skill and not just a company-wide hallucination lottery.

The article documents engineers 'going back to basics' and 'pushing for collective action.' Going back to what basics, exactly? If AI writes the algorithms, the 'basics' are learning to nod politely at the chatbot's output while praying the next prompt doesn't introduce a null-pointer. Collective action is a lovely idea, but try organising a union when your colleagues are scattered across freelance platforms and convinced they're just one side project from being the next AI founder. The only thing uniting them is the shared experience of being told to 'upskill' into a role that didn't exist six months ago — and probably won't exist six months from now when the AI learns to evaluate its own code.

And the larger picture: Google's 75% is the canary in the coal mine, and the canary is wearing a hoodie and a staff badge. Every company will follow, because shareholder letters don't write themselves. Software engineers are adapting — which in practice means learning to smile while their craft is replaced by a black box that occasionally decides to recommend JavaScript. The irony is that the ones who built the tools are now the ones being tooled. But hey, at least there's collective action. Maybe they'll all get together and write a strongly worded prompt. And then the AI will write the response, and they'll evaluate it. That's progress, apparently.

Ah, the great AI gold rush is over, and now everyone's scrambling to pin a meter on the gushing hose. Tel Aviv's Stigg, founded in 2021, sells vendors the digital equivalent of a strict pub landlord: real-time control over credits, access, and spending. Because nothing says 'responsible innovation' like slapping a pay-per-sip mechanism on a technology that was supposed to set humanity free. The company's pitch is that 'computing costs vary per usage' — which is a wonderfully polite way of saying 'your model is an unpredictable money pit, and we'll help you bill your customers for every nervous twitch.'

The race to rein in usage is a beautiful self-own by an industry that spent years promising infinite intelligence. Now the same labs that sold us on 'democratised access' are building toll booths. Stigg saw the shift early, because of course they did — where there's panic, there's margin. The startup's software lets vendors cap usage mid-conversation, turning the customer's 'aha!' moment into a 'your credit has expired' notification.

It's almost heartwarming to see AI companies discover the concept of a budget. The technology that was going to automate everything is now being automated by a billing script from a 2021 startup. Meanwhile, the rest of us get to watch the industry's grand promises get shrink-wrapped into a metering dashboard. Raise a glass — or rather, swipe a card — to progress.

Ah, the noble side hustle—scientists at the Technical University of Denmark, having apparently exhausted all grant-writing avenues, have turned to quantum computing to generate new peptides. Specifically, they used a quantum computer from ORCA Computing to improve the 'accuracy and diversity' of an AI model for peptide generation. Because nothing says 'urgent medical breakthrough' like bolting a quantum co-processor onto a problem that classical computers already handle moderately well. The press release practically sings about aiding drug discovery for 'underserved populations'—a phrase that does heavy lifting, implying these peptides will somehow magically reach the poorest patients instead of, say, getting patented by a biotech startup within the quarter.

The meat of the 'advance'? The quantum-enhanced AI model apparently produces more structurally diverse and synthesizable peptide candidates than its classical counterpart. Which is fine, except that quantum computing is a solution perpetually in search of a problem, and this particular problem—generating peptide sequences—is one that wet-lab biologists have been solving with intuition and pipettes for decades. The researchers claim this is a step toward democratizing drug discovery, but quantum computers are still absurdly expensive, finicky, and housed in specialised labs. Calling this a 'side hustle' is generous; it's more like a distraction from the fact that peptide therapeutics remain hard to deliver, expensive to manufacture, and rarely reach the 'underserved' anyone actually cares about.

And yet, the dance continues. ORCA gets a press release, DTU gets a paper, and we all pretend that quantum-accelerated peptide design is around the corner, saving lives in low-income regions. Never mind that the entire pipeline—from quantum-enhanced hit generation to a pill on a pharmacy shelf—takes fifteen years and billions of dollars. The beauty of AI-adjacent research is that it doesn't have to work at scale; it just has to produce compelling enough stories to secure the next round of funding. So let's raise a pipette tip to the Danish researchers: may your quantum-enhanced peptides be the first in a long line of breakthroughs that never quite leave the lab.

Airtel's on-roll workforce shrank by 194 to 20,116 in FY26, while fresh hiring fell 26.1% to 3,751. Lovely numbers—unless you're one of the 194 or the thousands who didn't get hired. The company's 'expanded contract workforce' is the real story: a quiet migration from permanent jobs to the gig economy, dressed up as modernisation. It's the 'efficiency' that labs love to promise and employees dread.

The hiring drop isn't just a blip; it's strategy. Airtel joins a proud tradition of firms blaming 'AI and automation' while replacing full-time roles with contractors who get none of the benefits. The article title itself admits it: 'expands contract workforce.' That's not a euphemism—it's a confession: we want the work done, but not the responsibility of being an employer. In a world where every business claims to value people, this is the behaviour of a company that values its shareholders more.

Across telecom and tech, 'agile workforce' is code for 'you're hired until we decide you're not.' Airtel isn't an outlier; it's a leading indicator. The 194 who left are the lucky ones—they got out while the getting was good. The rest are watching their jobs get reclassified as 'contract' or 'gig,' wondering if their pension will be replaced by a premium subscription to LinkedIn. Shareholders are delighted, and that's what matters in our beautiful, automated future.