ResidentialBusiness

After writing more than one article a day for the last 23 years, I’ve accumulated a body of text large enough to train an AI model that could convincingly write “like me.” With today’s technology, it would not be difficult to build a system capable of generating opinions that sound as if they came from Enrique Dans—an algorithmic professor that keeps publishing long after I’m gone. That, apparently, is the next frontier of productivity: the digital twin. Startups such as Viven and tools like Synthesia are building “AI clones” of employees and executives—trained on their voices, writing, decisions, and habits. The idea is seductive. Imagine scaling yourself infinitely: answering emails, recording videos, writing updates, etc., while you do something else, or nothing at all. But seductive doesn’t mean sensible. A world full of digital ghosts We are entering an era where professionals will not just automate tasks; they will replicate their personas. A company might build a digital copy of its best salesperson or customer service agent. A CEO might train a virtual twin to respond to inquiries. A university might deploy an AI version of a popular lecturer to deliver courses at scale. In theory, this sounds efficient. In practice, it invites a form of existential confusion: If the replica is convincing enough, what happens to the person? What does it mean to “be productive” when your digital version is the one doing the work? The fascination with cloning ourselves digitally reflects the same temptation that has driven automation for centuries: outsourcing not just labor, but also identity. The difference is that AI can now replicate the voice of that identity, both literally and metaphorically. What I would look like as an algorithm I could easily do it. Feed a large language model the millions of words I’ve written since 2003 — every article, every post, every comment — and you’d get a fairly accurate simulation of me. It would probably have the right tone, vocabulary, and rhythm. It could write plausible articles, maybe even publish them at the same pace. But it would just miss the point. I don’t write to fill a schedule or a database. I write to think or to teach. Writing, for me, is not an act of production, but of reflection. That’s why, as I explained recently in “Why I let AI help me think — but never think for me,” I never let AI write my articles for me. It makes no sense. Asking a model to think for me would defeat the very reason I sit down every morning to write. Of course, I use AI constantly: summarizing sources, checking facts, exploring counterarguments, and finding references. But I never let it finish my sentences. That’s the boundary that keeps my work mine. The illusion of scaling yourself The promise of digital clones is rooted in the same misconception: that replicating output equals replicating value. Companies now talk about “bottling expertise” or “scaling human capital” as if personality were a production line. But cloning output is not the same as extending competence. A person’s professional value is not their words or gestures. It’s their judgment, built over time through context and curiosity. A model trained on your past decisions may imitate your tone, but it cannot anticipate your evolution. It’s a fossil, not a future. An AI clone of me could mimic my writing style from 2025. But if I let it publish, it would freeze me in that year forever, a museum piece updated daily. From productivity to presence Executives, entrepreneurs, and creators should be careful what they wish for. A “digital twin” may handle the inbox or record video briefings, but it also dilutes what makes leadership or creativity meaningful: presence. In Axios’s coverage of CEO clones, many executives confessed that they liked their AI doubles but didn’t fully trust them. The clone could handle repetitive interactions, but not empathy, timing, or nuance—the qualities that define credibility. Delegating those things to an algorithm is like sending a mannequin to a meeting: technically present, emotionally vacant. Corporate immortality and the ethics of legacy There’s also the question of what happens when your digital twin outlives you. Some companies already treat employee data as assets, so why wouldn’t they treat their digital clones the same way? Imagine a firm continuing to deploy the “AI version” of a beloved leader or educator after they’ve passed away. It might seem like a tribute, but it’s really a form of corporate necromancy: using a person’s intellectual remains to perpetuate a brand. It’s not hard to picture universities selling “virtual professors” or corporations reusing former CEOs as permanent avatars. In a recent academic paper on digital twins, researchers warned that the boundary between “representation” and “possession” is getting blurry. Who owns the clone? Who profits from it? When we replicate people as data objects, we risk turning identity into infrastructure, into something that can be licensed, monetized, or rebranded at will. The right way to use AI for personal scale There is, however, a rational way to use AI for scale: as augmentation, not imitation. I use AI every day as a thinking partner. It reads drafts, proposes structures, suggests sources, and critiques my logic. It’s like having a tireless research assistant, one that never gets offended when I ignore its advice. But the act of reasoning, the decision of what to say and how to say it, remains mine. That’s the key difference between using artificial intelligence and becoming it. When we outsource thinking, we lose the feedback loop that makes us human: the constant process of reflection, revision, and growth. Professionals who embrace AI responsibly will amplify their reach without diluting their essence. Those who don’t will eventually find their own voices indistinguishable from their machines. What businesses should learn from this For companies flirting with employee clones or AI avatars, here’s a checklist worth remembering: Define purpose, not imitation. Don’t build AI twins to replicate people. Build systems that free them to do higher-value work. Keep the human in the authorship loop. AI can assist in drafting, coding, and summarizing, but final judgment must remain human. Treat data as legacy, not property. Respect employee and creator autonomy. No one should become a perpetual digital asset without consent. Focus on augmentation, not automation. Use AI to enhance collective intelligence, not to eliminate the need for it. AI is not here to replace human expertise; it’s here to challenge how we apply it. The paradox of self-replication Soon, anyone with enough data will be able to build a digital version of themselves. Some will see it as immortality; others, as redundancy. I see it as a mirror, a test of what truly matters in human work. When my own digital twin can write a decent article about AI, I won’t be impressed. The question isn’t whether it can write. It’s whether it can care, and whether it serves me for the purpose I’m trying to achieve. And until algorithms can care about truth, nuance, curiosity, or purpose, I’ll keep doing what I’ve done for the last 23 years: Sit down, think, and write. Not because I have to, but because I still can. View the full article

H Company and CEO Gautier Cloix turn AI and APIs into the next office colleague by creating agentic systems to perform real tasks alongside humans. View the full article

Do your office, inbox, and calendar feel like a ghost town on Friday afternoons? You’re not alone. I’m a labor economist who studies how technology and organizational change affect productivity and well-being. In a study published in an August 2025 working paper, I found that the way people allocate their time to work has changed profoundly since the COVID-19 pandemic began. For example, among professionals in occupations that can be done remotely, 35% to 40% worked remotely on Thursdays and Fridays in 2024, compared with only 15% in 2019. On Mondays, Tuesdays, and Wednesdays, nearly 30% worked remotely, versus 10% to 15% five years earlier. And white-collar employees have also become more likely to log off from work early on Fridays. They’re starting the weekend sooner than before the pandemic, whether while working at an office or remotely as the workweek comes to a close. Why is that happening? I suspect that remote work has diluted the barrier between the workweek and the weekend—especially when employees aren’t working at the office. The changing rhythm of work The American Time Use Survey, which the U.S. Labor Department’s Bureau of Labor Statistics conducts annually, asks thousands of Americans to recount how they spent the previous day, minute by minute. It tracks how long they spend working, commuting, doing housework, and caregiving. Because these diaries cover both weekdays and weekends, and include information about whether respondents could work remotely, this survey offers the most detailed picture available of how the rhythms of work and life are changing. This data also allows me to see where people conduct each activity, making it possible to estimate the share of time American professionals spend working from home. When I examined how the typical workday changed between 2019 and 2024, I saw dramatic shifts in where, when, and how people worked throughout that period. Millions of professionals who had never worked remotely suddenly did so full time at the height of the pandemic. Hybrid arrangements have since become common; many employees spend two or three days a week at home and the rest in the office. I found another change: From 2019 to 2024, the average number of minutes worked on Fridays fell by about 90 minutes in jobs that can be done from home. That change accounts for other factors, such as a professional’s age, education, and occupation. The decline for employees with jobs that are harder to do remotely was much smaller. Even if you just look at the raw data, U.S. employees with the potential to work remotely were working about 7½ hours per weekday on average in 2024, down about 13 minutes from 2019. These averages mask substantial variation between those with jobs that can more easily be done remotely and those who must report to the office most of the time. For example, among workers in the more remote-intensive jobs, they spent 7 hours, 6 minutes working on Fridays in 2024, but 8 hours, 24 minutes in 2019. That means I found, looking at the raw data, that Americans were working 78 fewer minutes on Fridays in 2024 than five years earlier. And controlling for other factors (e.g., demographics), this is actually an even larger 90-minute difference for employees who can do their jobs remotely. In contrast, those employees were working longer hours on Wednesdays. They worked 8 hours, 24 minutes on Wednesdays in 2024, half an hour more than the 7 hours, 54 minutes logged on that day of the week in 2019. Clearly, there’s a shift from some Friday hours, with employees making up the bulk of the difference on other weekdays. Fridays have long been a little different Although employees are shifting some of this skipped work time to other days of the week, most of the reduction—whether at the office or at home—has gone to leisure. To be sure, Fridays have always been a little different than other weekdays. Many bosses allowed their staff to dress more casually on Fridays and permitted people to depart early, long before the pandemic began. But the ability to work remotely has evidently amplified that tendency. This informal easing into the weekend, once confined to office norms, can be a morale booster. But as it has expanded, it’s become more individualized through remote and hybrid arrangements. Those workers in remote-intensive occupations who are single, young, or male reduced their working hours across the board the most, relative to 2019, although their time on the job increased a bit in 2024. The benefits and limits of flexibility There are a few causal studies on the effects of remote work on productivity and well-being in the workplace, including some in which I participated. A general takeaway is that people tend to spend less time collaborating and more time on independent tasks when they work remotely. That’s fine for some professions, but in roles that depend on frequent coordination, that pattern can complicate communication or weaken team cohesion. Colocation—being physically present with your colleagues—does matter for some types of tasks. But even if productivity doesn’t necessarily suffer, every hour of unscheduled, independent work can be an hour not spent in coordinated effort with colleagues. That means what happens when people clock out or log off early on a Friday—whether at home or at their office—depends on the nature of their work. In occupations that require continuous handoffs—such as journalism, healthcare, or customer service—staggered schedules can actually improve efficiency by spreading coverage across more hours in the day. But for employees in project-based or collaborative roles that depend on overlapping hours for brainstorming, review, or decision-making, uneven schedules can create friction. When colleagues are rarely online at the same time, small delays can compound and slow collective progress. The problem arises when flexible work becomes so individualized that it erodes shared rhythms altogether. The time-use data I analyzed suggests that remote-capable employees now spread their work more unevenly across the week, with less overlap in real time. Eventually, that can make it harder to sustain the informal interactions and team cohesion that once happened organically when everyone left the office together at the end of the week. As some of my other research has shown, that also can reduce job satisfaction and increase turnover in jobs requiring greater coordination. The future of work To be sure, allowing employees to do remote work and have some scheduling flexibility on any day of the week isn’t necessarily bad for business. The benefits—in terms of work-life balance, autonomy, recruitment, and reducing turnover—can be very real. Flexible and remote arrangements expand the pool of potential applicants by freeing employers from strict geographic limits. A company based in Chicago can now hire a software engineer in Boise or a designer in Atlanta without requiring relocation. This wider reach increases the supply of qualified candidates. It can—particularly in jobs requiring more coordination—also improve retention by allowing employees to adjust their work schedules around family or personal needs rather than having to choose between relocating or quitting. What’s more, many women who might have had to exit the labor force altogether when they became parents have been able to remain employed, at least on a part-time basis. But in my view, the erosion of Fridays may go beyond what began as an informal tradition—leaving the office early before the weekend begins. It is part of a broader shift toward individualized schedules that expand autonomy but reduce shared time for coordination. Christos Makridis is an associate research professor of information systems at Arizona State University, Institute for Humane Studies. This article is republished from The Conversation under a Creative Commons license. Read the original article. View the full article

Figure strengthens case for BoE rate cut next monthView the full article

It falls to Nvidia to make or break the market mood for the rest of the year View the full article

The most difficult problems can nurture the most talented researchersView the full article

Organisers of the tennis tournament warned that a price cap would damage its funding modelView the full article

Low allocation comes as the government and FCA encourage fund managers to put more money in domestic marketView the full article

Long neglected in Washington, the region is key to some of the president’s priorities, including halting illegal migration and limiting Chinese influenceView the full article

Test your knowledge of the set-piece political event with our questions on the financial statements of yesteryearView the full article

US claims meeting is part of an effort by Tehran to acquire sensitive military technologies from MoscowView the full article

Inspired by the ongoing auction of Bob Ross paintings to raise money for public television, Last Week Tonight With John Oliver is putting some of its own TV artifacts up for auction for a good cause. Host John Oliver dedicated the close of Sunday’s season finale to local public television, which is facing an unprecedented crisis. Federal budget cuts could by next year close as many as 115 public television and radio stations in the U.S. serving 43 million Americans, according to the Public Media Bridge Fund, a philanthropic initiative. “These stations can fill a vital community role,” Oliver said during Sunday’s show. johnoliversjunk.com Bob Ross Inc. said in October that it was putting 30 paintings by the late artist up for auction to pay for public station licensing fees. The first three paintings sold last week in Los Angeles for more than $600,000 total. Oliver said Last Week Tonight originally tried bidding on one of the recently auctioned Ross paintings in hopes of flipping it to raise even more money for public television. “Sadly, those prices were outside of our budget,” Oliver said. So instead, the show is tapping its own archives with the auction site johnoliversjunk.com. johnoliversjunk.com Items like the giant Reese’s mug that made its first appearance during a 2017 episode about net neutrality are now up for auction alongside Oliver’s “on-screen wife,” Mrs. Cabbage, and a quintet of bad wax replicas of presidents originally purchased by the show from the now-closed Hall of Presidents and First Ladies in Gettysburg, Pennsylvania. All the proceeds from the auctions will go to the Public Media Bridge Fund. johnoliversjunk.com Though Last Week Tonight didn’t have the budget to drop six figures on an original Bob Ross painting at last week’s auction, Bob Ross Inc. did donate one to Oliver’s auction. “Cabin at Sunset” was created during an 1987 episode of Ross’s PBS show The Joy of Painting, and it’s presently the first item shown on Oliver’s auction site. The painting currently has a bid of more than a million dollars. The top bid for a sculpture titled “LBJ’s Balls” is over $25,000, and the top bid for a trip to New York City to meet Oliver is higher than $50,000 at the time of this writing. So far, the leading bid to appear in a photo over Oliver’s shoulder during a future episode has just passed $100,000 after 45 bids. johnoliversjunk.com The show found some lower-priced ways to raise money, too, like signed merchandise from the Moon Mammoths, the minor league baseball team Last Week Tonight temporarily rebranded in July, and a Mr. Bean DVD signed by Joel McHale. The auction closes on November 24. Oliver also promoted Adopt A Station, a nonprofit for people who want to help out and donate to public media stations but aren’t able to participate in his auction. The President administration budget cuts meant an end to the Corporation for Public Broadcasting (CPB), which said in August that it is winding down operations. The Public Media Bridge Fund says the end of CPB funding will destabilize the public media system. It’s seeking to raise $100 million over two years to help the most at-risk communities. View the full article

The influence of the AI industry is becoming a major topic in New York’s 12th congressional district, where a crowded Democratic primary packed with millennial and Gen Z candidates is heating up. The seat represents one of the wealthiest communities in the country — and is a liberal stronghold — so whoever wins could eventually become a major player in the fight to limit the most noxious impacts of large language model (LLM) technology. On Tuesday, Cameron Kasky, a political activist and Parkland shooting survivor who lives in the district (which includes the Upper West Side and Upper East Side) announced he was running. His campaign is making fighting the “AI oligarchs” a pivotal focus, adapting the rally cry used by progressive Democrats like Bernie Sanders and Alexandria Ocasio-Cortez, for the ChatGPT age. “Generative AI is undoubtedly one of the most societally damaging innovations that humanity has ever created, and people do not understand the toll it will be taking on us,” says Kasky’s new campaign site. “This damage includes the fresh water supplies it is depleting, a media literacy crisis that has already gotten out of control in this country over recent years, and the degree to which children are leaning on AI for therapy, companionship, and more — at the cost of their critical thinking skills and cognitive development.” Kasky says his legislative priorities will include holding AI companies accountable for their environmental impact, preventing mass layoffs, and better regulating the influence of tech companies on child safety. “I have no sympathy for AI, and no tolerance for what it has done to our population. It will only get worse if we do not get in the way as aggressively as possible,” he says. He’s not the only person planning to take on AI in the primary. Alex Bores, a Palantir alum who has proposed state legislation – the RAISE Act – that would rein in the industry, has also made clear that one of his focuses would be regulating artificial intelligence. Earlier this week, he was targeted by ads funded by Leading the Future, a pro-AI super PAC funded by OpenAI executive Greg Brockman and Andreessen Horowitz that’s intent on blunting the influence of tech critics in the upcoming congressional primaries. The group has called Bores’ legislation a “clear example of the patchwork, uninformed, and bureaucratic state laws that would slow American progress and open the door for China to win the global race for AI leadership.” Bores, in turn, has started fundraising off the ads. “This AI super PAC’s first target? Me,” said Bores in a tweet. “Why? They’re scared of leaders who understand their business regulating their business. They want unchecked power at your expense—and I’m the guy standing in their way. It is a crowded race, with about ten people running for the Democratic nomination in total. Most of the other candidates, who include Nadler favorite Micah Lasher, community organizer Liam Elkind, and attorney Jami Floyd, have yet to issue strong positions on artificial intelligence – but that could change. Meanwhile, Kennedy family heir and social media provocateur Jack Schlossberg, who also announced his campaign this month, has at least some thoughts on the tech. Last year, he tweeted his reflections: “Question about AI — Is it sexual ? We’re are ALL sexual beings, that’s just a fact. If AI is non-sexual, does that limit its potential ? or make it unstoppable ?” View the full article

Billionaire’s appearance at black-tie event for Mohammed bin Salman signals improving relations with Donald The PresidentView the full article

US president touts $1tn investment from Mohammed bin Salman during gilded Oval Office welcome for strongmanView the full article

Competition and Markets Authority set to probe dentistry sector after sharp increase in pricesView the full article

Across the country, data center demand and construction have been skyrocketing throughout 2025. And so has local opposition to those projects. From Indiana (where a developer withdrew its application to build a data center on more than 700 acres of farmland after local opposition) to Georgia (where now at least eight municipalities have passed moratoriums on data center development), residents and politicians are pushing back against the water- and energy-hungry sites. Between late March through June of this year alone, 20 data center projects, representing about $98 billion in investments, were blocked or delayed in the United States, according to a new report from Data Center Watch, a project from the AI security and intelligence firm 10a Labs. That number is higher than all of the data center disruptions the research group had tracked in the two years prior to its most recent report. A turning point against data centers Data Center Watch began keeping tabs on this trend in 2023, and released its first report earlier this year, covering 2023 through the first quarter of 2025. In that time frame, 16 projects, worth $64 billion, were blocked or delayed. Though a project may be cancelled for myriad reasons, these were cases where local opposition was reported to have played some role in the decision, says Miquel Vila, an analyst at the Data Center Watch project. In the second quarter of 2025, that opposition surged 125%. “We were expecting a few more cases,” Vila says of Q2, “but not 20.” One important caveat, Vila notes, is that the data center industry is booming; it makes sense that opposition would, too. But even accounting for record high construction spend, he sees these recent numbers as a “turning point” in the trend. What’s wrong with data centers? Tech giants are building out data centers at a rapid pace to meet the enormous power needs of artificial intelligence (AI). But data centers have faced local criticism because of the resources they consume, like water (which is especially a concern in scarce regions like Arizona) and energy (which has been linked to rising electricity prices across the country.) Along with water use and utility prices, communities have also taken issue with noise, landmark preservation, and transparency, Vila adds—like if it isn’t clear who the end user of a data center will be. Data Center Watch has found 188 community groups that have formed to fight data center projects. Between March and June alone, 53 active groups across 17 states were targeting 30 data center projects. Amid that pushback, lawmakers have also been reconsidering their regions’ tax subsidies to data centers, as well as regulations around zoning, and the projects’ environmental impacts. That community opposition is even causing some lawmakers to change their regulations or hold off on building data centers in the future. “Local opposition is having an impact in the regulatory landscape of data centers,” Vila says. Dan Diorio, vice president of state policy for the industry group Data Center Coalition, said in a statement that it continues to see “significant interest” across the country for “responsible data center projects,” and said such projects create jobs, economic investment, and local tax revenue. He added that the coalition’s members are committed to community engagement, stakeholder education, and to working with policymakers and regulatory bodies. “Data centers are also committed to being responsible and responsive neighbors in the communities where they operate,” Diorio said. Data centers and politics Data center opposition has become a talking point in recent political races. In Virginia—the biggest data center market in the world—Governor-elect Abigail Spanberger campaigned in part on making sure data centers pay “their fair share,” and on addressing rising electricity prices. In Georgia, Peter Hubbard—who was elected to the state’s Public Service Commission, which regulates its utilities—has specifically highlighted how data centers can drive up people’s energy bills. Georgia is increasingly becoming a data center hotbed, and is in fact the second-largest such market in the world. But while both those politicians are Democrats, data center opposition is a bipartisan issue, Data Center Watch found. Both blue and red states are rethinking incentives to developers or tightening their rules around such projects. That tracks with other research about data center support: a recent Heatmap poll found that only 44% of Americans would welcome a data center near them. Looking ahead Data Center Watch plans to keep an eye on project delays and cancellations going forward. Already, it seems the trend is continuing into Q3: In one prominent example, Amazon’s proposed Project Blue data center, was rejected by Tucson, Arizona’s town council in August. (In Data Center Watch’s latest report, two of the 20 affected projects were from Amazon: one in Becker, Minnesota, which was suspended as lawmakers reconsidered tax incentives, and one in King George, Virginia, which was delayed because of legal issues and resident pushback.) Vila expects data center opposition to keep growing—and to increasingly become a part of project calculations. “Before, local opposition was more of an anecdotal possibility,” he says. “Now, it’s becoming a core feature of development . . . in the same way issues like land, energy, and water are taken into account.” View the full article

Pigs famously have thick skin and Donald The President does not. It’s just one of myriad distinctions between the cloven-hoofed barnyard animal and America’s 47th president. There’s a good reason, however, why many social media users are currently addressing The President as “Piggy,” and sharing crude, AI-assisted images of him in porcine form. Rest assured, he paved his own pathway to hog heaven. On Monday, a clip of The President addressing reporters aboard Air Force One went viral. It begins with reporter Jennifer Jacobs pressing The President about the eternally unfurling Epstein scandal. The president seems as though he’d rather not answer the question—at least, that’s how it comes across when he admonishes Jacobs: “Quiet, Piggy.” While leaders in most professions might be disciplined or even fired for such a transgression, The President has proven uniquely immune to formal consequences for violating norms. But he is in no way immune to informal consequences, which is why the internet has already repurposed ‘Quiet, Piggy’ into a memetic insult against The President. Bluesky users have started quote-tweeting The President’s latest TruthSocial dispatches with the new catchphrase, and they’re doing the same for media appearances from The Presidentian underlings like House Speaker Mike Johnson and U.S. Representative Nancy Mace. Quiet, Piggy. — Kevin M. Kruse (@kevinmkruse.bsky.social) 2025-11-18T13:30:52.321Z “Quiet Piggy.” 🙄 — Amanda Weaver (@amandaweavernovels.com) 2025-11-18T16:50:48.163Z Over on X, Governor Gavin Newsom is among the many users adding a body-shaming component to the catchphrase, tweeting unflattering photos of the president along with it. Quiet, piggy. pic.twitter.com/RIKsI4iDjV — Gavin Newsom (@GavinNewsom) November 18, 2025 “Quiet, piggy.” pic.twitter.com/3uOoRnjGpX — Rick Wilson (@TheRickWilson) November 18, 2025 Quiet, piggy pic.twitter.com/CKORVzGGpG — Republicans against The President (@RpsAgainstThe President) November 18, 2025 Meanwhile, some TikTok users are also posting unflattering images of the president to accompany the insult, and others are posting AI-generated images of the president, alternately as Miss Piggy or as himself yelling at Miss Piggy. If social media users seem especially eager to weaponize “Quiet, Piggy” by reflecting it back at the president, it’s likely because of how well this outburst fits in with The President’s previous behavior. The President has a documented history of calling women like Rosie O’Donnell and former Miss Universe Alicia Machado “pigs”—along with “dogs,” “slobs” and “disgusting animals”. He also has a more recent and pointed history of insulting and berating journalists. Just after the 2016 election, 60 Minutes journalist Lesley Stahl reportedly said that The President told her the reason he regularly bashes reporters is to “demean” and “discredit” them so that the public will not believe “negative stories” about him. And The President continued to insult journalists in his tone-setting first post-election press conference, refusing to take a question from CNN reporter Jim Acosta and telling him: “You are fake news.” Over the course of his initial term, The President would escalate attacks on press that seemed to be insufficiently friendly, deeming them the “enemy of the people.” He seemed to harbor a special animosity, though, toward journalists who happened to be women. In one typically fiery exchange with CNN’s Abby Phillip in 2018, for instance, The President responded to Phillip’s question about then-Special Prosecutor Robert Mueller by saying: “What a stupid question that is. What a stupid question. But I watch you a lot, you ask a lot of stupid questions.” In his second term, The President appears even more committed to attacking reporters for asking questions he’d prefer not receive. He regularly refuses to answer questions, tells reporters “You’re not supposed to be asking that,” or calls them “obnoxious” and “very evil” for asking anyway. Indeed, the whole “TACO The President” attack over the summer, which accused the president of Always Chickening Out on tariffs, would likely not have blown up to the level it did had The President not told a reporter who asked him about it: “Don’t ever say what you said.” Still, despite The President having been extra combative with reporters all year, he has lately seemed even more prickly with an uptick in questions about his connection to Jeffrey Epstein. The President on Epstein Files: I don’t want to talk about it because fake news like you—you’re a terrible reporter—fake news like you just keeps bringing up to deflect from the tremendous success of The The President Admin pic.twitter.com/rZjobTujCN — Acyn (@Acyn) November 16, 2025 The President: Will you let me finish? You are the worst. You’re with Bloomberg right? You are the worst. I don’t know why they even have you. pic.twitter.com/mTmZ77KTYv — Acyn (@Acyn) November 17, 2025 When an ABC reporter asked The President about the Epstein files on Tuesday, during the course of this writing, The President responded by saying, “I think the license should be taken away from ABC,” and urging FCC chairman Brendan Carr to “look at that.” As heated as The President can get when asked about this issue, though, “Quiet, Piggy” stands out as an exceedingly juvenile and degrading insult. Many social media users have been speculating about why the schoolyard name-calling went unchallenged in the moment; why Jacobs’s fellow reporters didn’t make sure her question got answered or demand an apology on her behalf. Perhaps it’s the absence of any heroes aboard Air Force One, though, that has inspired social media users to push back on The President’s hogwash themselves. View the full article

A "spike in unusual traffic" caused service degradation for the infrastructure giant, disrupting digital banking for customers. View the full article

The billionaire and legacy government-sponsored enterprise investor says there is a quick interim fix and they should eventually leave conservatorship. View the full article

This post is part of Find Your Fit Tech, Lifehacker's fitness wearables buying guide. I'm asking the tough questions about whether wearables can really improve your health, how to find the right one for you, and how to make the most of the data wearables can offer. When I test smartwatches and fitness trackers, I always pay attention to accuracy. Is my running pace correct? Does the device capture the ups and downs of my heart rate? I even got a VO2max lab test to check a bunch of watches' fitness scores. But you'll notice that one thing I don't test for accuracy is calorie burn. None of my devices come close to agreeing on the number of calories they think I'm burning, and I don't expect them to. Even scientists who study the accuracy of wearables can't answer the question in a way that's useful when you're shopping for this year's devices—but we'll get into why that is below. There was a time, before Fitbits, when nobody knew quite how many calories they were burning on a daily basis. Sure, you could calculate a rough estimate based on your body size, sex, and age. You could choose whether or not to believe the calorie readout on the cardio machines at the gym. (Spoiler: don't.) But the idea that a gadget on your wrist could tell you how many calories you personally burned during one particular day was revolutionary. It was also wrong. How fitness trackers calculate calorie burnBefore I discuss how accurate fitness trackers are, let’s look at where they get their numbers. For calorie calculations, the main sources are motion and heart rate data. For motion, trackers use accelerometers to figure out when your body is moving, and by how much. If you have a watch on your wrist, and the watch swings back and forth rhythmically while sort of bouncing up and down, your gadget guesses that you must be walking. If there is quicker bouncing and your wrist makes a smaller movement, you’re probably running. This is the basic idea behind how trackers detect how many steps you’re taking. If you’ve paid attention to your step count, you already know some of the ways this can be inaccurate. If you’re shopping, for example, keeping your hand on the shopping cart handle may result in you not getting credit for the steps you’re taking. That depends on the device, though. (For a perfect illustration of this issue, see these tests I did comparing a Garmin to an Apple Watch on a treadmill. When I rested my hands on the treadmill handle, the Apple Watch recorded 318 steps while the Garmin recorded none.) Then there’s the heart rate sensor: Since your hands don’t always move predictably during exercise, it can be easier to just tell your watch that you’ll be cycling or doing yoga or whatever. The gadget then uses your heart rate to make an educated guess about how much work your body is doing. Whatever the source of the data—heart rate, movements, or a combination—the gadget processes it through a formula to calculate how many calories it thinks you’re burning. Your age, weight, and sex may figure into this equation. Generally, though, the fitness tracker doesn’t actually know how many calories you’re burning; instead, it’s calculating a probable number based on incomplete information. Why there's no simple test for accuracyIf humans were robots, all built the same, all moving in predictable patterns, this formulaic approach might work. But humans are complicated, and technology often gets confused. For example, you may get different step counts if you put a device on your right versus left wrist. And the optical heart rate sensors that a lot of trackers use may be less accurate on dark skin compared to lighter skin. These problems relate to the data that the trackers gather, but calorie burn isn't a direct measurement. It's a calculation, and different algorithms can come up with different calorie burn numbers depending on how the algorithm is designed. The companies that make fitness trackers aren’t required to publish their algorithms or verify that their calorie counts are accurate. They can just put a device on the market, and there you are, comparing wearables on shopping sites without any information about how accurate they are, outside of the companies’ claims. Researchers are interested in fitness trackers’ accuracy, which would seem like a good thing. They want to be able to use wearables in research or recommend them for individuals and healthcare providers, and so they'll run studies comparing consumer devices to lab equipment. This sounds like a great way to answer our questions! But there’s a huge delay in actually getting that information, and it’s often published too late to be useful. By the time a researcher buys a batch of the latest model, runs a study, writes it up, submits it to a journal, and finally gets it published, several years may have gone by, and the company has moved on to the next model. That delay is why I (usually) can't use scientific studies to weigh in on the devices I write about. Here's a great example of how frustrating it can be: this review was published in 2025, and found the Series 1 was the Apple Watch that turned up the most often in the studies the authors were able to gather. The Series 9 and 10 watches were completely missing from the available data, and as a reminder, we're now up to a Series 11. With that caveat about delays, I still think it’s useful to look at the research on fitness trackers to see what themes emerge. Are any of them good at estimating your calorie burn? What studies say about fitness trackers’ accuracyTime for the bad news. A study from 2020 that looked at a variety of gadgets from Apple, Garmin, Polar, and Fitbit found that all the devices are inaccurate more often than they are accurate. The authors considered a device to be accurate if its reading was plus or minus 3% when compared to a more reliable measure of energy expenditure (that is, calorie burn) in a lab setting. Here’s how some of the top brands fared: Garmins underestimated calorie burn 69% of the time. Apple watches overestimated calorie burn 58% of the time. Polar devices overestimated calorie burn 69% of the time. Fitbits underestimated calorie burn 48% of the time and overestimated 39% of the time. The fact that Fitbits were roughly correct on average doesn’t mean they were useful. If sometimes your device overestimates and sometimes it underestimates, it’s not very helpful unless you know which is which. A 2018 review specifically of Fitbits found that accuracy varied greatly depending on factors like where they were worn (torso was more accurate than wrist), whether you were walking uphill, and whether you walked at a constant speed or stopped and started. The accuracy also varied by device, with the Fitbit Classic underestimating calorie burn and the Fitbit Charge usually overestimating. The devices just aren’t accurate enough to know how many calories you’re really burning A 2022 study compared the Apple Watch 6, the Fitbit Sense, and the Polar Vantage V. The researchers had volunteers wear all three gadgets while sitting quietly, walking, running, cycling, and strength training. Every gadget, for every activity, was awarded a judgment of “poor accuracy,” with coefficients of variation ranging from 15% to 30%. And this 2025 review of Apple Watch studies—the same one I mentioned above—found that calorie burn was, on average, off by about 18%. The authors mention a proposed standard from the International Electrotechnical Commission that recommends fitness trackers not be off by more than 10%. That's clearly not being met, although the researchers noted that newer models may be more accurate than earlier ones. To get a sense of these percentages, let's say your true calorie burn is 2,000 calories per day. A device that's off by 15% might report that you burned 1,700 calories, or that you burned 2,300. If you're using your device to figure out how much to eat, you could be way off in meeting your calorie goals. If these devices are all inaccurate, how can you know how many calories you're burning?It’s probably most useful if you think of your calorie burn as a number you cannot measure directly. Treat it as a black box: I burn some unknowable number of calories, now what? The only common reason you would need an accurate estimate of calorie burn is if you are trying to figure out how much food you need to eat. If you want to lose weight, you want to eat less than you burn; if you want to gain weight, you want the reverse; and if you’re trying to maintain your weight, you want to eat roughly the same as what you burn. But think about it this way: you don't actually need to know your calorie burn if you have the other two terms in the equation—your calorie intake, and your weight. It's considered more accurate to adjust how much to eat based directly on your weight, rather than using calorie burn estimates as a middleman. Let’s say you’re training for a marathon and you want to make sure you fuel yourself appropriately. Well, if you’re under-eating, you’ll start to lose weight. When you start to see the scale trending downward, that’s your signal to add a few hundred calories to your diet. If, after that adjustment, your weight stays steady, then you know you’re eating the right amount. As you increase your training (or if you take time off to rest a sprained ankle), you can make more adjustments as you go. I have a post here detailing how to make these adjustments with the help of either a paid app, a group of free apps, or a DIY spreadsheet. If you’ve been using a fitness tracker instead, and it’s working for you, feel free to keep using it. But if the tracker ever stops giving you the results you want, you can safely leave it out of the equation. View the full article

The multi-year, $100 million agreement will allow users to take financial actions without leaving the ChatGPT app. View the full article

The average loss from the two categories is almost seven times higher than the mean amount for all other types, according to research from ALTA and Milliman. View the full article

We may earn a commission from links on this page. Deal pricing and availability subject to change after time of publication. It can be tricky to find reliable Bluetooth earbuds that hit the right balance of affordability and quality (especially ones with ANC), but the JBL Tune Buds deliver. They’re currently the lowest price they’ve ever been, according to price-trackers, at $39.95 (originally $99.95), bringing the original price down 60%. JBL Tune Buds $39.95 at Amazon $99.95 Save $60.00 Get Deal Get Deal $39.95 at Amazon $99.95 Save $60.00 They’ve earned an “Excellent” rating from PCMag, which commends their decent noise cancellation, customizable EQ, and built-in Alexa, which are all the more impressive at a sub-$50 price point. The earbuds feature JBL’s classic bass-heavy sound, and with the customizable EQ in the companion app, you can further tweak the audio to your liking. The three pairs of silicone eartips the buds ship with also allow users to adjust fit and comfort. While they can’t compare to top-tier noise cancellation from brands like Bose and Beats, they block outside noise fairly well and last up to 10 hours per charge (12 hours with ANC off). This gets boosted to up to 36 hours with the charging case, which makes up for its somewhat bulky silhouette. They support Bluetooth 5.3 and have 10mm dynamic drivers that deliver a powerful sound with plenty of bass. They’re slightly more durable than similar noise-cancelling buds thanks to an IP54 rating, which can withstand light rain and sweat. If you’re looking for a pair of earbuds with modest noise cancellation, customizable and punchy sound, and long-lasting battery life for the price, the JBL Tune Buds are solid everyday earbuds that perform well and won’t break the bank—especially at the current 60% discount. Our Best Editor-Vetted Tech Deals Right Now Apple AirPods 4 Wireless Earbuds — $117.00 (List Price $129.00) Apple iPad 11" 128GB A16 WiFi Tablet (Blue, 2025) — $299.00 (List Price $349.00) Shark AV2501AE AI XL Hepa- Safe Self-Emptying Base Robot Vacuum — $294.99 (List Price $649.99) Amazon Fire HD 10 (2023) — $69.99 (List Price $139.99) Sony WH-1000XM5 — $248.00 (List Price $399.99) Blink Outdoor 4 1080p Wireless Security Camera (5-Pack) — $159.99 (List Price $399.99) Ring Floodlight Cam Wired Plus 1080p Security Camera (White) — $99.99 (List Price $179.99) Amazon Fire TV Stick 4K Plus — $24.99 (List Price $49.99) NEW Bose Quiet Comfort Ultra Wireless Noise Cancelling Headphones — $298.00 (List Price $429.00) Deals are selected by our commerce team View the full article

Microsoft said Tuesday it is partnering with artificial intelligence company Anthropic and chipmaker Nvidia as part of a cloud infrastructure deal that moves the software giant further away from its longtime alliance with OpenAI. Anthropic, maker of the chatbot Claude that competes with OpenAI’s ChatGPT, said it is committed to buying $30 billion in computing capacity from Microsoft’s Azure cloud computing platform. As part of the partnership, Nvidia will also invest up to $10 billion in Anthropic, and Microsoft will invest up to $5 billion in the San Francisco-based startup. The joint announcements by CEOs Dario Amodei of Anthropic, Satya Nadella of Microsoft, and Jensen Huang of Nvidia came just ahead of the opening of Microsoft’s annual Ignite developer conference. “This is all about deepening our commitment to bringing the best infrastructure, model choice and applications to our customers,” Nadella said on a video call with the other two executives, adding that it builds on the “critical” partnership Microsoft still has with OpenAI. Microsoft was, until earlier this year, the exclusive cloud provider for OpenAI and made the technology behind ChatGPT the foundation for its own AI assistant, Copilot. But the two companies moved farther apart and their business agreements were amended as OpenAI increasingly sought to secure its own cloud capacity through big deals with Oracle, SoftBank, and other data center developers and chipmakers. Asked in September if OpenAI could do more with those new computing partnerships than it could with Microsoft, OpenAI CEO Sam Altman told The Associated Press his company was “severely limited for the value we can offer to people.” At the same time, Microsoft holds a roughly 27% stake in the new for-profit corporation that OpenAI, founded as a nonprofit, is forming to advance its commercial ambitions as the world’s most valuable startup. Anthropic, founded by ex-OpenAI leaders in 2021, said Claude will now be the “only frontier model” available to customers of the three biggest cloud computing providers: Amazon, which remains Anthropic’s primary cloud provider, and Google and Microsoft. AI products like Claude and ChatGPT take huge amounts of energy and computing power to build and operate, and neither OpenAI nor Anthropic is yet turning a profit. As part of the deal, Nvidia said Anthropic will have access to up to a gigawatt of capacity from its specialized AI chips. Huang said he’s “admired the work of Anthropic and Dario for a long time, and this is the first time we are going to deeply partner with Anthropic to accelerate Claude.” —Matt O’Brien, AP technology writer View the full article