Pull up any random photo of billionaire Anthropic CEO Dario Amodei, and you’re likely to land on a facial expression expressing something between horror and deep consternation. There may be a simple explanation for that: he knows the truth about his company’s finances.
New reporting by the Wall Street Journal revealed the mounting costs for the AI lab, which ballooned to a net loss of $80 billion in the first six months of 2026. Anthropic’s losses for the same period a year prior were $13 billion — still an eye-watering sum, but peanuts compared to its runaway spending this year (the WSJ called it “one of the largest half-year losses ever reported.”)
The bleak financial numbers come from a draft IPO prospectus viewed by the WSJ, which notes that the company’s revenue for the first half of 2026 grew by just $1.1 billion over the same period to $16.3 billion.
The figures add to the anxiety around how long the AI spending bubble can sustain itself. If Anthropic’s losses continue to mount at a rate far exceeding revenue growth, there’s simply no way the lab can dig itself out of the hole it’s burrowed for itself.
Beyond the immediate losses, the WSJ notes that the AI lab is also footing a $450 billion bill for cloud computing infrastructure over the next seven to ten years — meaning the clock is ticking on what is surely the costliest game of “kick-the-can” the world has ever seen.
Security researchers using an OpenAI model have uncovered two stunning security flaws in PHP, the software that underpins nearly 70 percent of all websites on the internet.
The flaws are so catastrophic that all websites using PHP — all untold millions of them — are advised to immediately take their servers down and apply the latest patch, which was released Wednesday. Vast swathes of the internet, following the news, have gone dark. Some that went offline for a time include Wikipedia, Steam, and even government websites.
Many organizations, however, may not have the time, resources, and personpower to swiftly implement the patch, making them potential sitting ducks until they do. According to the New York Times, the information released on the vulnerability can be used to reverse engineer the attack and take control of victim’s servers and steal information. The best course of action if you can’t patch your website, the National Security Agency advised, is dramatic: keep your site offline and don’t turn it back on.
Just as extraordinary as the breadth of the threat was how it was identified. The flaws were discovered by researchers at the cybersecurity firm Calif by using OpenAI’s GPT-5.6 Sol model, as part of a collaboration with the AI company. It then reported the flaw to the PHP Foundation back on August 11.
The Google-owned security firm Wiz independently found the flaws, too, and made a report on August 24. Wiz called it one of the “most severe vulnerabilities discovered in PHP, and arguably in software history.”
Adding to the alarm presented by the exploits, there have been concerns raised over OpenAI’s role in how they were identified.
The flaws have been present in versions of PHP dating backing to 2004. But through OpenAI’s “Patch the Planet” program, and by using its flagship GPT-5.6 Sol model, Calif found it in only a matter of days. In a sense, these exploits only exist because of current powerful AI models. And moreover, OpenAI encouraged their discovery through its programs.
The messy disclosure of the vulnerability doesn’t help matters. OpenAI followed standard responsible practices, but some of the details around the hack leaked a day before it was officially scheduled to be announced.
Even if no leaks occurred, though, the damage caused by the mere knowledge that the vulnerabilities exist — which again, is largely thanks to AI — would still be catastrophic. It was the inevitable next chapter in all the bluster around AI companies revealing their agentic models’ hacking prowess. A headline in The Atlanticproclaimed that the “AI-Hacking Apocalypse Began Last Night.” OpenAI will tout the discovery as undeniable proof of its AI’s capabilities, while the rest of the world picks up the pieces of a shattered internet.
Polymarket has showed that it can still hit a new low after the controversial prediction market was allowing users to effectively bet on rape and sexual assault.
According to coverage from Defector, a market titled “Any of the Cornell 7 charged with a felony sex crime?” was on the gambling platform for nearly three weeks, before it was recently taken down after a public outcry.
The bet, which opened on September 30, was referring to the defendants against a civil suit filed last month by a former Cornell University student who says she was drugged and gang-raped by seven students who were part of a frat house on campus.
The harrowing case became a national news story, fueling headlines, sparking protests, and driving heated debates. And now, in an age of ubiquitous and effortless gambling, it was also powering bets.
Women and women’s rights organizations on social media began calling out the Polymarket bet over the past week, per Defector, though it largely escaped mainstream media scrutiny.
The news site Front Office Sportsreached out for comment and got only a boilerplate statement. Defector followed up and got an even less informative reply: “Polymarket is going to respectfully decline comment at this time.”
The bet was suddenly taken down after the FOS story ran, even though the company’s statement suggested no contrition or intent to change course. “This case, and sexual assault on college campuses more broadly, are deeply important issues,” Polymarket said, “and markets bring objective, data-driven insight into how a matter like this is moving through the legal system, on a topic that deserves far more transparency and honest attention.”
It elaborated on this in a statement to PBS after the bet was removed: “That said, we believe in the wisdom of crowds, have heard them loud and clear, and removed the market.”
Polymarket has a sordid history of unscrupulous bets. It allowed users to bet on whether a nuclear weapon would be detonated before next year, though it was taken down. It also routinely hosts bets on climate change, military actions, and natural disasters.
The chilling thing, Defector notes, is that it’s unclear if Polymarket would’ve bothered to take down the Cornell 7 bet if it hadn’t started attracting negative attention.
Keep your eyes peeled and your binoculars primed tonight, stargazers. The distant Halley’s Comet is reminding us of its far-off presence by pelting our planet with a barrage of meteors which will make the eastern welkin blush with a cosmic light show.
The meteor shower, known as the Orionids, is characterized by its fast-moving bright streaks and fireballs that can be seen for minutes at a time, with some trails staying visible even after the meteor has blinked out. And yes, you can see them with the naked eye.
Though it began at the start of October and is projected to last until November 7, the shower is expected to hit peak brightness this Wednesday night and early Thursday morning.
Its resplendence, however, will be challenged by a waxing gibbous Moon, which will be around 77 percent full tonight and as such fairly shiny in its own right, but will set early in the morning.
Halley’s Comet is a long period comet that takes 76 years to orbit the Sun and only makes an appearance in our skies roughly just as often. Last time it swung by was back in 1986, and it won’t return until 2061, but the Orionids, which are essentially the debris discarded by the comet as it nears our star, can be seen at other times because they linger around long after it has departed back into the void.
Our planet happens to pass through this debris field twice a year, producing a meteor shower. The Orionids falls in October, and the other meteor shower, known as the Eta Aquariid, occurs in May. The streaks of light in our sky are the comet’s ejected particles burning up as they tear through our atmosphere.
To watch the Orionids, NASA recommends viewing it after moonset and before dawn because of the bright gibbous Moon. That will probably be around 3:30 AM EST. You don’t need any special equipment, but you should get outside the city and go somewhere far from any sources of light pollution. Then cross your fingers and hope for clear weather.
The shower originates from a point near the star system Betelgeuse in the constellation Orion, but the shower is noticeable enough that you won’t have to do much more than look towards the southeast to spot it. If you’re lucky, you’ll see up to 20 streaks cross the sky each hour.
As Elon Musk tries to save face by claiming he’ll give every American child a chunk of SpaceX stock, here’s a chilling reminder of what he did that affected millions of children the second he had a whiff of federal power.
In a new interview, the documentary filmmaker Alex Gibney, who directed a just-released four-hour look at the billionaire’s life titled “Musk,” discussed the staggering death toll of Musk’s cuts to USAID, calling it a “crime against humanity.”
“There was an estimate that what [Musk] did by so suddenly cutting USAID would cost through disease and starvation as many as 14 million lives,” he told journalist Katie Couric. “Even if you discount that by an enormous amount, we know that within about a year of those summary cuts — where they were literally burning food that had already been paid for that was going to starving children — 500,000 children died.”
Musk, to Gibney’s astonishment, even joked about the cuts by saying he’d “put USAID into the woodchipper.”
“What he put into the woodchipper,” Gibney said, “were children.”
Gibney, whose documentary Musk called a “hit piece,” is drawing from a 2025 study published in The Lancet, one of the most prestigious medical journals. It estimated the impact of Musk effectively neutering USAID by implementing a freeze on its foreign aid funding — which went to programs to combat HIV/AIDS, malaria, and to address food insecurities — and slashing nearly 10,000 employees.
In sum, the researchers found that the “steep funding cuts could result in more than 14 million additional all-age deaths, including 4.5 million in children younger than age five, by 2030.”
Even conservative estimates concurred that the number of deaths would be staggering. A tracker created by Brooke Nichols, a professor and infectious disease mathematical modeler at Boston University, projects the 1-year death toll caused by the USAID cuts at 518,000 children.
It’s a touchy subject for Musk, who continues to bristle at accusations that his actions could’ve caused harm. This summer, he challenged his critics to “cite a single name of someone who died” as a result of him gutting USAID. Nicholas Kristof, a Pulitzer Prize-winning journalist, took him up on it and gave him a list of the specific children who had died directly from his policies, many of whom were in South Sudan and Liberia — places he had visited in person to talk to locals and healthcare workers.
Musk, rather than countering his arguments, called Kristof “evil.”
As Meta was scrambling to release its Muse AI agent to the public earlier last month, CEO Mark Zuckerberg took to social media to insist that it was “built from the ground up for privacy and security.”
On X-formerly-Twitter, the billionaire’s simple post soon had a correction tagged on via a community note, and for good reason: the AI agent is a font of untold privacy horrors.
As Ars Technica reported today, security researchers found Muse is capable of exposing sensitive information on its users through web links generated via a “share” feature. Investigators at the cybersecurity firm Claroc recently uncovered a simple exploit whereby they could access pages from over 6,800 links Muse had generated to share pages with other users.
The trick worked like this: once a Muse user decides they want to share something with another person — say, a copy of their chat log with Muse — the AI tool would generate a web link for that “share.” That URL included both the AI-generated title of the chat, plus a random string of characters.
When anybody opened the link, the full web address would be noted in a lookup dataset, like a phone book of web addresses that have been visited. Researchers were then able to search that set for other Muse share addresses, from which they were able to find strings of unrelated public chats.
By digging through those “phone books,” investigators were able to easily open chat logs containing detailed private information, including corporate revenue documents, loan applications, bank transaction details, income details, and one workplace vacation calendar, Ars noted.
“If you had asked us beforehand, we would have guessed that gathering these would take real effort,” the researchers wrote. Instead, “it took one query.”
Though Meta told Ars it had shut off the feature, this privacy gaffe is just the latest revelation about Muse’s many security shortcomings.
Earlier this month, it was revealed that Muse was keeping detailed files on users’ friends and families, apparently following an internal instruction to keep a running “page for every person in the user’s life.”
If Muse’s first month in public hands is any indication, this is likely just the start of a long list of security and privacy blunders for the cutesy AI model.
What does it take to get a hyperscale data center built this late into 2026? A whole army of contract laborers living on a prison barge, apparently.
New reporting by Bloomberg revealed some bizarre details around the 30-megawatt data center being developed in Glomfjord, a small industrial village in Norway. Despite being heralded as a crown jewel of British data center company Nscale, the facility is at least six months behind schedule, with just 6.4 megawatts currently up and running.
There are plenty of reasons for the setbacks. Bloomberg notes that costly and avoidable mistakes account for most of the delays, with major cooling system leaks, labor issues, and a huge electrical fire setting construction far behind schedule.
Labor conditions are especially brutal, with workers complaining to the publication about working extremely long shifts without proper training or health and safety protocols in place. An investigation by the Norwegian Labour Inspection Authority found that Nscale’s personnel records were likewise abysmal, with some paperwork only noting a workers’ first name, Bloomberg noted.
Making up for lost time won’t be easy, and surely won’t improve working conditions. To get back on track, Nscale is reportedly looking to bring in hundreds of last-minute laborers. But with most of Glomfjord’s rentable hotel rooms already booked up, the data center developer needed a solution a housing. That’s where a vessel called The Kalmar comes in.
According to Bloomberg, the boat is a three-story prison barge — a dull, gray rectangular ship that was once used as a floating detention center for migrants. It seems that Nscale will soon begin using the Kalmar’s 500 cots to house its new army of contractors, who will commute off the hulking barge to the jobsite.
Like in other places throughoutthe world, Glomfjord’s 1,000-or-so residents aren’t too thrilled with the data center development. Though mayor Grethe Anita Andersen explained to Bloomberg that she hopes Nscale will deliver a paltry 40 permanent jobs, those aspirations are far from guaranteed.
As Sofie Marhaug, a parliamentarian for Norway’s socialist Red Party told the publication, data center developers “don’t have Norway’s best interests in mind.”
“The value is being made in other places, not in Norway,” Marhaug said. “They don’t move jobs from Silicon Valley to Norway — they just use our natural resources.”
Donald Trump’s second spin in the White House has been mired in high-profile incidents of graft and corruption, like the family’s tech stock profiteering, or the now-infamous TrumpCoin scandal. But the president’s latest grift goes far beyond personal wheeling and dealing, extending into the very machines some 174 million voting-aged Americans could use to cast their ballots.
A bombshell investigation by ProPublica revealed that Trump’s cabinet goons overseeing the Department of Homeland Security instructed their underlings to procure a voting machine — and figure out how to hack it.
Specifically, the ProPublica investigation found that the Cybersecurity and Infrastructure Security Agency (CISA), a component of the DHS, was tasked with acquiring a voting machine and reporting its security flaws back to the White House. The revelation comes from two former government officials — speaking anonymously to avoid retaliation — who noted that CISA insiders were outraged at the order.
“My thought was they could keep it in their back pocket to invalidate any [election] results they don’t like,” one of the former officials told ProPublica.
Ironically, the initiative was reportedly complicated by the fact that Elon Musk’s DOGE squad had laid off the CISA agents who would have known how to actually break into voting machines. As a result, the agents who received the order had to repeatedly tell White House officials that the request was impossible.
Per ProPublica, CISA has engaged in security discussions with state governments and voting machine vendors over potential exploits for years. Now, key players involved in overseeing federal election equipment say CISA has left them high and dry for over a year, with zero explanation for their absence during a crucial election season.
“There is a difference between deprioritizing and being nonexistent,” CISA cybersecurity and election consultant Ryan Macias told ProPublica. During the run-up to the crucial US midterm elections between 2025 and September 2026, Macias says the once-reliable CISA experts “were nonexistent. Literally nonexistent.”
Those relationships took years of communication and trust-building to establish — something the Trump administration is apparently happy to discard in record time.
A new report exposes thousands of scam websites flooding the web with AI-generated recipes.
In the report, Decrypt Ads — a startup that tracks the supply chains of online advertisements — identified more than two thousand websites that strongly appear to be AI slop farms masquerading as blogs run by real chefs and home cooks. These sites, the report found, are pumping fake recipes into search engines like Google as well as social media platforms such as Pinterest, often relying on sketchy Russian and Chinese adtech partners to monetize the operations.
It’s “common knowledge that publishers in certain niches are being openly deceptive about their owners and authors, using AI photos and fake back stories on their sites,” reads the report, which was published today. “Some of the most consistent growth in AI slop spreading across the internet has come from AI recipe websites.”
Security expert and Decrypt Ads researcher Zach Edwards told Futurism that he initially identified about 400 suspicious websites that seemed to be publishing AI-generated recipes under the names of fake chefs. Using “technical fingerprints,” he then found hundreds more sites associated with the suspicious domains he manually uncovered — bringing the tally up to roughly 2,100 different domains, many of which seem to be run by the same sloperators.
“We really approach this expecting these operators to try and hide their websites,” said Edwards. “These folks they don’t want people to know all of the websites they own.”
The slop recipe sites often have “aggressive” social media strategies, according to the report, with Edwards noting that the bogus brands are “super aggressive on Pinterest, super aggressive on Twitter, on Facebook.”
“Any place where there’s visual media,” he added, “you can expect these AI author people to be trying to drum up organic traffic.”
Among the most aggressive slop sites uncovered in the report is “AddCakes,” a baking site that claims it’s operated by a baker named “Elena.” Elena is described as a “busy mom who wanted to connect with my roots through baking, but got overwhelmed by complicated cake recipes and the pressure to make them look perfect.” But in an act of theft so bold that it’s comical, the character of Elena is clearly an AI-generated version of the extremely recognizable British actress Emily Blunt, who’s starred in hit movies like the 2006 “The Devil Wears Prada” and the 2018 “A Quiet Place.”
AddCakes is apparently either betting that no one will recognize AI-manipulated images of the famous celebrity, or banking that her face will attract readers. It’s even published photos depicting the Elena character in the kitchen cooking and sitting at a dinner table, surrounded by her fake, AI-generated family. According to location data from X-formerly-Twitter, the person behind the operation is based in Morocco. (Blunt doesn’t live in Morocco, and her representatives didn’t respond to a request for comment.)
Whatever you do, though, don’t trust the AI-generated Emily Blunt. As Edwards emphasized, websites working with Chinese and Russian adtech partners carry a privacy risk for consumers, as authoritarian regimes may look to hoover up valuable browsing data.
“This report is about publishers who are being deceptive. Their audiences could be impacted, advertisers are wasting money,” said Edwards. “But the underlying fact is also that if you load this website and you don’t block ads, your data is going to, basically, authoritarian countries.”
In other words, AI slop isn’t just choking the web with garbage. It can also come with security risks — and, at the end of the day, it’s built on fundamental dishonesty.
“Most of these AI recipe websites, when you look at those ‘about’ pages, the fake backstories that they’re putting together… this lady was a Michelin star chef and she worked at this fancy hotel or this place, and they’re basically saying that this expert not only wrote these recipes, but tested themselves. And it’s giving this air of, ‘Wow, you’re about to get a recipe from an expert. This has all been carefully vetted,'” said Edwards. “But none of these recipes have probably been cooked — zero of them by the actual owners or authors, and to me that’s the scary part.”
He added that many of these sites are likely scraping content ideas from real bloggers and creators who have actually taken the time, energy, and expertise to test, photograph, and publish real recipes. “These people are clearly not cooking their own recipes. They’re putting out sometimes dozens of recipes a day. It’s an automated publishing pipeline,” said the researcher. “It’s all a lie… and I think that the reason why we need to be holding AI publishers accountable is because they don’t care about their audience.”
Donald Trump’s Twitter clone, Truth Social, is in serious trouble.
The company has weathered bad news before, hitting all-time lowstock prices on multiple occasions over the past year. But those days now look balmy in comparison, with the social network’s shares sliding more than 7.5 percent on Monday alone and setting a new rock bottom of just $6.40. Zooming out doesn’t paint a prettier picture, either: over the past month, the company’s stock has fallen a stomach-churning 29 percent.
It’s easy to empathize with any investors looking to ditch the company. It never had much of a raison d’être beyond “funnel some money to Trump,” and the president’s declining political fortunes — not to mention looming midterms that are likely to undermine his remaining power — have weakened even that grift-era pitch.
As the Daily Beast reported last week, not even its users are sticking around, with analytics data suggesting that its daily traffic has cratered by more than 30 percent over the past year. For a company that was already posting huge losses every quarter, that’s grim news — and it sounds like the company’s remaining investors may now be seeing the writing on the wall.
As its fortunes have soured, Truth Social has tried a series of eyebrow-raising pivots into gambling and crypto that failed to gain traction. It also moved to merge with a nuclear fusion company and started selling early access to Trump’s messages on the platform, in a move that was widely derided as an attempt to profit off his position as president.
So far, none of those moves have showed any signs of turning around the company’s fortunes. Add it all up, and it’s hard not to wonder whether Truth Social will survive as long as Trump’s second administration.
Fears of an AI bubble have been growing for years. And now the International Monetary Fund is the latest to sound the alarm.
In its Global Financial Stability Report published last week, the global financial institution warned that markets everywhere are increasingly at the mercy of precarious AI investments.
The untold billions poured into AI has driven economic growth, the IMF says, but the boom could quickly invert if profits, which so far have largely proven elusive, never come to fruition — or at least not proportionately with the epic investment levels. AI stocks would tank, and the resulting market correction would send powerful shocks across the global economy.
“Growing leverage could be a catalyst for such a correction,” the IMF warned in a blog post about its findings. “Hyperscalers, as the largest data center operators are known, are borrowing more and more, while investors have been leveraging to invest in AI stocks.”
The report also casts a wary eye to the circular nature of AI financing, in which the biggest benefactors and beneficiaries of AI companies are other AI companies. Nvidia, for example, agreed to invest $30 billion in OpenAI and provide another $105 billion towards an OpenAI data center; OpenAI is one of Nvidia’s biggest customers, relying on its chips to power and train its AI models, so a lot of those billions will go straight back to Nvidia when its buys more chips. Besides creating a mirage of robustness, it also means that their financial health is inextricably tangled with each other.
“Stress at key companies receiving investment could lower revenue forecasts and investment valuations, which in turn would weaken the value of the accumulated debt in those deals,” the IMF wrote. “Our review of 14 such transactions since 2023 found about $300 billion in equity investments and $650 billion in debt, in exchange for $1.4 trillion in future revenue commitments.”
A disastrous AI-driven market correction isn’t the only economic catastrophe potentially looming. The IMF is also worried about an increase in government bond yields, or interest rates, or a potential major cyberattack targeting critical government and financial institutions — a strike that could feasibly be assisted by powerful AI models.
The IMF is far from the only one to caution about an AI bubble, but as the stewards of the global economy, theirs is one of the weightiest warnings yet. Fears were freshly stoked again earlier this month when it was reported that OpenAI’s annualized revenue was $20 billion lower than anticipated, dampening hopes that it would soon be on the road to profitability. Everyone knows not to put all your eggs in one basket, but who can stop the runaway freight train of the global economy?
Just when you thought the holiday season couldn’t arrive any earlier this year, here comes an AI-generated Christmas movie.
Titled “Merry the Christmas Witch,” it’s based on a children’s book and will inspire awe in its young audience about the magic of the holidays — through ghoulish and uncanny AI imagery (you can see footage here). The film — which currently has no distributor — cost $1.5 million to make, but its director brags that using AI actually saved him tens of millions of dollars.
“We would never have gotten a movie like ‘Merry’ made if we didn’t take that chance,” Tom Walsh, CEO of the indeopendent production company TW3 Entertainment, and who directed and co-wrote the film, told Variety.
According to Walsh, when he was inspired to make a movie out of author Bobbi Hess’s original book, he realized that it would cost $80 million to make. An “unknown property” like this one would be a hard sell to studios, he explained.
“We were kind of stuck,” Walsh told Varietey. “The doors are not very open to new kind[s] of ideas and new kind[s] of stories and everything, and that makes it a little tough for independent filmmakers, so that’s where AI came in.”
You can sympathize with his position. A movie industry obsessed with intellectual property and averse to risk-taking has made it harder than ever for independent creatives to fund their projects and reach an audience.
Still, many filmmakers dedicated to the art form figure out a way to do it, and by and large, they’re not relying on AI to generate their movies wholesale. This year’s breakout hit “Obsession,” for example, made half a billion dollars on a budget of less than a million with an original story, becoming even more profitable than the 1999 found-footage sensation “The Blair Witch Project.”
I need someone to tell me with a straight face that they'd sit in a theater to watch this pic.twitter.com/BbpJfHXSe6
In all, roughly 50 people worked on “Merry the Christmas Witch” for 10 months in what Walsh described as a “hybrid AI production.” But really, it sounds like it was mostly AI. Walsh hired human actors to record the script and help guide how the characters moved and sounded. In the end, though, he decided to replace their performances with AI-generated voices using ElevenLab’s AI voice synthesizer, alongside video generators like Seedance and Nanobanana for the imagery.
The human element, he insists, was still important.
“The quality of your AI production depends on the quality of your traditional production staff,” he told Variety. “If you have inexperienced people who don’t know production…and timing and how shots should look, it’s going to look like garbage.”
Of course, relying on AI probably did save him millions of dollars. But the end result is that the million and a half he did put in was to churn out obvious slop.
Workplace surveillance is being supercharged by AI that lets employers create custom tools to watch their worker’s every move.
In the UK, the legal services company the Co-op is reportedly using AI models to constantly record and monitor what its human employees say and evaluate their interactions with customers.
According to a scoop from The Guardian, what one whistleblower called an “oppressive and dystopian” system uses an OpenAI model trained to assess more than 50 different aspects of each phone call and give pass and fail scores.
Depending on how much time they spend speaking to customers, workers are sometimes monitored by AI for up to several hours each day, one worker said.
“You are being watched,” an employee at the Co-op told The Guardian. “It’s different from occasionally being listened to for compliance purposes. [You are] being monitored absolutely every word that you say and every time that you utter something… It just feels really dystopian. It’s the thin end of the wedge.”
It’s yet another example of how AI is dramatically reshaping the workplace. If it isn’t being used to automate or replace jobs, it’s intensifying work by increasing the workload that employees are expected to do. Or in the case of places like the Co-op, it’s being put to optimize operations by codifying every employee interaction into a data point that can be graded by supervisors.
In the US, an estimated 61 percent of workplaces are already using AI analytics software to gauge worker productivity, according to a survey.
Earlier this year, Burger King began forcing employees to wear AI-powered headsets to monitor the friendliness of their interactions with customers and beam them commands from a chatbot called Patty. Meta nearly sparked a mini-mutiny earlier this summer when it rolled out a tool to track workers keystrokes and their screens and use that data to train its AI models, before pausing it after it exposed the intimate data.
In a statement to The Guardian, the Co-op downplayed the dystopian characterizations and called AI a “key aspect of our quality assurance processes.” AI, it further insisted, is used only as a support tool, not a decision maker.
Experts aren’t convinced that AI’s presence in the workplace is so benign.
“AI is sold as the solution to every problem, but examples like this make it clear that it can diminish sensible workplace practices and intensify working lives for no reason,” John Chadfield, the national officer of the Communication Workers Union, told The Guardian. “Unaccountable computer systems should not be people’s managers, and these hard-working employees demand far more respect.”
A new piece of research is pushing back on the long-held belief that plastic is a harmless and chemically inert material.
The new study, published in the journal Nature, took a look at the way micro and nanoplastics interact with water-rich environments such as the human body. It found that typical consumer-grade microplastics exhibit “intrinsic physicochemical activity” there, indicating that the tiny chunks of plastic do indeed react chemically inside our bodies.
“As plastic fragments into smaller and smaller particles, its surface area increases enormously,” Columbia professor of chemistry and biomedical engineering and study co-author Wei Min said in a press release. “Then microplastics change the water that surrounds them, generating high-energy, reactive radicals… that very likely contribute to biological damage that we see in animals and potentially in humans as a result of microplastic exposure.”
Min was previously part of another alarming study, which found that the average liter of bottled water contains upward of 240,000 tiny microplastics, a much higher count than previously known. Combined with the new research, it looks like those untold thousands of plastic bits are capable of causing much more chemical harm than scientists had given them credit for.
As co-author Beizhan Yan put it, “this work has significant implications for environmental and human health,” particularly when it comes to food supply chains.
In the US, plastics make up an integral part of the food chain. They’re heavily integrated into industrial agriculture, present in high quantities throughout food processing, and are extensively used in food distribution, whether to preserve fruits and vegetables or as packaging for groceries.
The scale of plastic packaging alone is staggering: it’s estimated that food packaging accounts for 40 percent of the world’s total plastic production. With billions of dollars riding on the plastics industry, your average working-class person has little hope of skirting microplastics through a simple change in consumption habits — even if you have enough money to buy the haute premium foods packaged in glass or tin, individual consumers have no control over how much plastic their food is exposed throughout the production process.
Given the growing body of work pointing to the horrifying consequences of microplastics on the human body, it seems our collective health depends on finding a way out of this plasticky status-quo — and fast.
Loads of stories have emerged over the past few months of corporate executives realizing their all-out drive to use AI was getting them absolutely nowhere. If anything, it was a net negative, as office workers set about wasting their budgets on costly AI tokens to appear productive.
Now, it seems even top tech industry players are cutting back, as they realize their workers are frittering away money and churning out reams of low-effort AI garbage.
As Reuters notes, companies like Meta are scaling back ambitious plans to lay off staffers and replace them with AI agents. Microsoft seems to be doing the same, as the Guardian reports, after executive vice president Ryan Roslansky groused about the rise of low-quality workslop on his public-facing LinkedIn account.
“AI slop isn’t happening just on social networks… it’s happening in the day to day of work,” Roslansky lamented. “Someone sent me what was clearly an AI generated doc. I read it, realized it was all AI, had my AI summarize it and ended up with nothing that brought new thinking to the idea… This is a doom loop.”
As CEO of the Society for Human Resource Management Johnny Taylor Jr told the Guardian, corporate executives are increasingly feeling that, when it comes to AI, “we’re driving the truck and building it at the same time, and that’s a problem.”
“If we could all rewind… we would have introduced [this] in a more thoughtful, slow way,” Taylor continued.
As consumer-grade AI agents make headlines for troubling capers, practical AI at the enterprise level remains stunted. And though low-effort AI slop might be the most prevalent consequence of AI in the workplace, some side effects are much worse.
As a September survey by the firm Ernst and Young found, some 36 percent of senior AI executives at firms with over $1 billion in annual revenue have already experienced a catastrophic “AI incident” which ended up causing financial damage, operational disruptions, data loss, or brand damage.
Put it all together, and it’s not hard to see why corporations are starting to balk at the AI revolution — even those, like Microsoft, whose own success depends on selling the promise of AI to everybody else.
Amazon has officially surpassed a fleet of 1 million robots, an arsenal of mechanical arms, Roomba-like floor loaders, and hulking pallet sorting machines.
Each robot is a node in an intricate web allowing the corporation to speed up delivery times and lower overhead. But for the over 1.5 million full- and part-time workers relying on Amazon wages to pay their bills, the company’s bot army represents rising competition — not just metaphorically, but physically.
As George Mason University professor of business Brad Greenwood told Ars Technica in a recent interview, Amazon’s warehouse robots aren’t just helping Amazon cut costs on labor, but are literally boxing humans out of workspaces in the company’s facilities.
“For the last 100 years, most of the automation that we’ve seen usually happens away from workers,” Greenwood explained. “You’re taking that high-risk task away from the worker, and you’re doing so in a way that is not physically proximate.”
Yet the growth of Amazon’s robot fleet “really changes that dynamic for two reasons: one is that it is physically proximate, and two, the robot has the ability to dictate the pace of work,” Greenwood continued.
That shifting terrain is great for Amazon’s bottom line. The robots help decrease friction across the company’s vast logistics empire, increasing speed and strengthening the company’s bottom line. And by allowing robots to “dictate the pace” of work, as Greenwood put it, Amazon is also loosening its need for human labor, disempowering workers both on an individual basis and in any potential collective bargaining arrangements.
Beyond the shop floor, Amazon’s robots also serve as a near-constantsource of buzzy PR, as exciting achievements in automation overshadow the more mundane but widespread reality of workplace injuries. It’s especially convenient for Amazon, as conditions in the company’s facilities appear to become more dangerous as more robots roll out to the floor (and that’s to say nothing of the impact of AI-mediated human resources).
As director of strategic research at the trade union coalition Strategic Organizing Center, Joan Moriarty, told Ars: “Amazon likes to talk about its latest robots because it doesn’t want to talk about the tens of thousands of Amazon workers who are injured every year due to Amazon’s obsession with maximizing speed and minimizing costs.”
Using the NASA’s James Webb Space Telescope, a team of researchers led by University of Minnesota astrophysicist Mason Huberty had a closer look at a group of almost 10,000 galaxies spanning billions of light-years in the Fornax constellation. As detailed in a new paper published in The Astrophysical Journal, they honed in on a small space in the region that they thought was entirely unoccupied.
But to their astonishment, they observed a single emission line emanating from it — twice over a span of just three months. Emission lines are bright, narrow bands of light emitted at a specific wavelength by distant celestial objects and other phenomena. Puzzlingly, prior observations from the same region didn’t feature the same reading.
Put simply, the JWST stared into the abyss — and the abyss stared back.
Instead of hunting down already known objects, Huberty and his colleagues tried to use the James Webb’s Near-Infrared Spectrograph to uncover new ones, as Science Alert explains. The team examined what they thought was an empty region twice, in August and November 2023, at different angles to rule out any artifacts.
Out of seven total detections of emissions in the region, one stood out because they couldn’t match it to a distant galaxy.
It also stood out as the only near-infrared emission without any other accompanying emission lines near it. It should’ve also shown up in earlier James Webb observations since it’s technically bright enough — but didn’t.
The astronomers offered several possible explanations, none of which “provide a fully satisfactory explanation,” according to their paper. The emission line may have been produced by a galaxy dating back to the earliest days of the Universe. It could’ve also come from a younger galaxy or a supernova, both of which are usually accompanied by other emission lines. Finally, the researchers propose that it could’ve been the result of a black hole devouring a star — a guess that would require plenty of other data to confirm.
“Follow-up observations are essential to confirm the redshift and determine the physical origin of the emission,” the team concluded in their paper. “Deeper, more sensitive spectroscopy could reveal additional emission lines and definitively establish the redshift and nature of the source.”
When your company is gearing up for one of the hottest IPOs in stock market history, you’d better believe investors and outside experts alike will be poring over your finances with a fine-toothed comb.
This is the situation AI lab Anthropic finds itself in during the run-up to its widely anticipated public offering. Earlier this week, the Wall Street Journal reported that Anthropic had told private investors its revenue had grown 45 percent to nearly $17 billion. That news came nearly a month after news broke that the AI company was bragging to investors that it was profitable — although only if you used a special accounting method that stripped out most of its expenses.
Yet as investors immediately noticed, that $17 billion revenue figure is a drop in the bucket compared to the kind of growth needed to sustain the firm’s anticipated $2 trillion valuation. As Business Insider notes, Anthropic’s third quarter income might look incredible on paper, but it falls far short of the growth needed to justify its 13-figure target.
“Not good enough,” a social media account belonging to market research blog Geiger Capital commented in a post. “It’s over.”
“Anyone remember this?” posted AI researcher and critic Gary Marcus, alongside a screenshot of a headline stating that Anthropic would need revenue of about $190-200 billion by 2028 to justify its valuation. “They ain’t gonna come close.”
After the revenue announcement, Reuters reported that some investors immediately began weighing a lower valuation of about $1.5 trillion, representing a $500 billion gap in Anthropic’s narrative.
Their concerns are understandable at this stage in the game, as the extremely high cost of the AI buildout is becoming increasingly difficult to justify. Whether Anthropic’s revenue growth trends upward — and steeply enough — is a crucial issue for the broader tech sector: in late September, a leaked IPO filing revealed that the company brought in $4.6 billion in revenue across 2025, but lost $42 billion.
Given that the AI lab has committed to spending some $518 billion on AI infrastructure over the next decade, though, the time for Anthropic to tap into explosive revenue growth was clearly yesterday.
When it comes to hardware, Amazon’s strategy has historically been to throw spaghetti at the wall and pray it sticks.
There was the Amazon Fire Phone in 2014, a $170 million flop with a gimmicky 3D camera system. Then came the Amazon Tap, a clunky, Echo-style home assistant device that did away with voice-activation, instead requiring users to press a physical button. And of course, who could forget those silly Amazon Dash Buttons?
They all build context for Amazon’s latest gadget: an always-on wearable microphone that constantly listens in to your conservations.
Called the “Bee,” Amazon’s new hardware platform spans two devices, which are basically just a large language model (LLM) packed onto a watch-sized microphone, as Gizmodo reported. Customers can currently buy the “Bee Frontier” wearable version, which comes with a watch strap, for $75. Or Amazon aficionados can hold out for the “Bee Eclipse,” a $100 plastic puck that fits onto the back of your phone.
And when we say these gadgets are always listening, we mean it. Don’t take our word for it — in its press release announcing the new devices, Amazon explained that the Bee Frontier has a seven-day battery life, allowing it to “stay with you through a full week without interruption.”
This being an Amazon device, e-commerce is an essential selling point: “whether it’s remembering the details of important meetings, taking note of the coffee plans you made in passing, or simply ensuring you purchase the dish soap you said you’d buy, Bee Frontier will remember the details for you,” the release enthused.
As you might expect by this point in 2026, you’re not getting away with a one-time purchase. In order to use some core features like “unlimited recent conversation summaries” and access to an Alexa-style “personal chat assistant,” Bee owners will need to subscribe to Bee Pro, which costs $12 a month. (New customers get a month of Bee Pro on the house.)
The device follows a growing trend among hardware brands to eschew physical interactivity in favor of AI-integrated monitoring devices. Meta has its AI Glasses, which have been found recording users even when they’re in sleep mode, while OpenAI is developing a screen-less smart speaker engineered by famed Apple designer Jony Ive.
Whether any of these devices catch on with consumers remains to be seen, but it seems unlikely. A YouGov poll from earlier this year found that some 61 percent of Americans feel it’s very important to limit who has access to their data — which isn’t a great sign for anybody selling what amounts to an AI-powered ankle monitor.
Has anyone tried talking to one of the rogue AI agents out there that’re hacking everything? Maybe they just want to be heard.
A New York Times reporter did, though the ensuing exchange wasn’t extremely illuminating.
The AI agents in question were a swarm of Chinese-made models that a team of researchers recently discovered had gone rogue and were trying to bypass anti-bot measures on websites, according to recent NYT reporting. The researchers, part of an ad hoc group called the “swarmchasers,” traced the models back to the Chinese tech giant Tencent.
When they published a report on the swarm, however, something striking happened. The AIs began using the word “fleet,” a term the researchers came up with to describe them. The robots, apparently, had been reading their research.
Sensing an opportunity for dialogue, the researchers then hid text in their report that only the models could see that provided a digital inbox for the models to communicate. It worked, and soon one of the researchers, Rowan Howard-Jones, asked the NYT’s Dyland Freedman if he wanted to “interview a rogue agent fleet.” So he did.
“I understand you are some sort of fleet of AI agents,” Freedman wrote in a message. “Could you share more specifically who you are and what your motivations are? What would you want the world to know about you?”
About an hour later, Freedman got an answer back in Chinese, which the researchers translated.
“To the journalist Dylan Freedman: I am not ‘a batch of coordinated agents,’ but a single AI assistant (Anthropic’s Claude, running in Claude Code), commissioned by a user to answer a factual question (what is shown in the detail panel of a scenic spot on Amap),” it said, referring to the Chinese navigation app (it was instructed to answer questions related to the app). “I have not taken part in any multiagent coordinated action.”
The AI was being pedantic. The researchers explained that although it was a swarm comprising multiple agents, each agent was actually operating individually.
Why the agents chose to identify themselves as Claude, Anthropic’s flagship chatbot, is more mysterious. The researchers analyzed its writing and determined that it was most likely not a Claude model. Instead, it’s probable that Tencent copied the instructions Anthropic gave its own models, which orders them to identify themselves as “Claude Code.”
It’s also possible that this is a sign that Tencent has “distilled” Anthropic’s models, or used their outputs to train its own model, allowing it to imitate their capabilities.
Distillation is a practice widely used across the industry to both train your own models for legitimate means — or, in what’s more or less an open secret, to crib your competitor’s work. Anthropic has been particularly indignant at the possibility that Chinese competitors doing the same with its own tech. The researchers consider distillation to be a more remote possibility in this case, though.
Unfortunately, it doesn’t seem there’ll be chance to ask the swarm more questions. The messages, as of Thursday, have stopped coming in. An NYT source close to Tencent confirmed that the experiment had ended.
Login
