All Activity

Google updated its Ads Transparency policy to show more detailed information about the entities funding ads – a move aimed at increasing accountability in digital advertising. Driving the news. Starting today, Google is displaying the payer name – drawn from an advertiser’s payment profile – if it differs from their verified advertiser name. For agency accounts, the client’s payment profile will be used as the payer name when applicable. This information will appear in the My Ad Center panel and the Ads Transparency Center. Why we care. This update increases public visibility into who is actually funding ads, which can impact brand perception, trust, and compliance. For agencies and businesses managing multiple clients or brands, accurate payer naming is essential to avoid misattribution, confusion, or reputational risk. It also introduces new operational steps – like re-verification and profile name management – that advertisers need to handle to maintain transparency and control how they’re represented. The next phase. Starting in June, advertisers can manually edit their payer name in their account’s verification settings. New Google Ads accounts will default to using the payment profile name unless it is updated during verification. Between the lines. This change aims to prevent confusion and ensure the true source of ad funding is clear, particularly in cases where agencies buy ads on behalf of clients. Google advised agency advertisers who were misclassified as direct advertisers to re-verify their accounts by May 31 to avoid inaccurate payer name displays. Zom out. Election Ads advertisers are already required to disclose payer names during verification. Those wishing to update their payer name must re-complete the election ad verification process under existing policies. The announcement. Updates to Ads Transparency Policy (April 2025) View the full article

Users are visiting Google more often but staying less. That’s one of several insights from a new analysis by Kevin Indig, in partnership with Similarweb, based on 5 billion search queries and 20 million websites across the U.S., UK, and Germany. The big picture. In the U.S., visits to Google are up +9% since AI Overviews launched in May 2024. Also: Time on site and pages per visit are flat or falling across markets. Query length has barely changed. The numbers point to a new user pattern: more frequent visits, shorter sessions – a shift toward “resolve and leave” behavior, according to Indig. By the numbers. Google visits are up in the U.S.: U.S. visits rose from 26.9% to 29.1% after the launch of AI Overviews last May. Pageviews on websites from AI Overview keywords have grown 22%. Meanwhile, engagement is falling: Time-on-site for Google.com is flat or declining in all three markets. In Germany, where AI Overviews launched two months ago, time on site dropped noticeably. Pages-per-visit dropped post-launch, suggesting users found answers faster. Also, query length is mostly unchanged: U.S. average query length rose just +3% since January 2023, from 3.27 to 3.37 words. In the UK, query length declined slightly after AI Overviews rolled out. Germany saw a modest +0.4% increase – but that was before AI Overviews went live. Why we care. Google AI Overviews are now seen by more than 1.5 billion searchers every month and occupy the most valuable real estate on Google’s search results. Google owns what searchers see (and don’t see), so SEOs and brands must understand how this impacts their behavior. The report. The Data Behind Google’s AI Overviews: What Sundar Pichai Won’t Tell You by Kevin Indig View the full article

McDonalds released its earnings report for the first quarter of 2025 early this morning. According to the report, U.S. comparable sales decreased by 3.6% from the previous year, primarily driven by reduced guest counts. This is the fast food giant’s largest U.S. sales drop since the second quarter of 2020, when U.S. sales plunged 8.7% during the height of the COVID-19 pandemic. ‘Grappling with uncertainty’ The fast food retailer reported that U.S. store revenue dipped to $5.96 billion, below a Bloomberg consensus estimate of $6.12 billion cited by Yahoo Finance. In the earnings report, McDonald’s CEO Chris Kempczinski noted that “Consumers today are grappling with uncertainty.” Chicago-based McDonald’s said adjusted earnings per share were $2.67. The earnings report also indicated that global sales fell 1.0%, with the most notable decline in sales in the U.K. McDonald’s shares (NYSE: MCD) were down 1.53% in early-afternoon trading on Thursday following the report. The stock is up 7.71% year to date. Fast food embracing value meals amid consumer caution Consumers appear to be tightening their spending due to economic uncertainty, likely not helped by President The President’s broad tariffs and erratic trade policies, which have made people anxiousness about increased costs and a possible recession. In this morning’s earnings call, Kempczinski warned, “we’re not immune to the volatility in the industry or the pressures that our consumers are facing.” McDonald’s menu prices have notably risen along with inflation in recent years, leading to customer backlash. More recently, however, the fast food retailer has attempted to increase customer traffic by releasing new menu items and promoting value-focused deals. Recent discount deals include the McValue menu, which features buy one, add one for $1 items, and $5 Meal Deals. Other fast-food chains have made similar moves to boost sales. Meanwhile, the burger giant plans to continue to offer meal deals like this, according to McDonald’s CFO Ian Borden. “While we may adjust our current McValue offerings over time,” Borden said on the earnings call, “for the remainder of 2025, we’ll continue to include everyday value meal deals starting at $5 given how the current $5 meal deal in particular has resonated with customers.” View the full article

We may earn a commission from links on this page. Though the marketing was a bit cagey, it's clear by this point that Ryan Coogler's latest, the critical and box office powerhouse Sinners, throws vampires into its genre-busting mix unashamedly. Set largely in 1932, the movie follows a couple of twin brothers (both played by Michael B. Jordan) returning to their backwater hometown of to Clarksdale, Mississippi to set up what they hope will be a lucrative juke joint—but not before navigating old relationships and the Klan leader who owns the sawmill that the two are looking to buy. By the midway point, all the movie's more grounded elements feed into a larger narrative blending magical realism, action, and horror. It's a thoroughly distinctive vision, a movie that's made a great deal of money, and an early Oscar favorite—so look for studios to learn the wrong lessons and give us a stream of generic vampire movies. Not here, though! None of these movies is particularly like Sinners, other than the vampire angle—but just as Ryan Coogler's movie uses vampires as a starting point, these movies all use undead lore in wildly different ways to tell wildly different, but consistently good, stories. Ganja & Hess (1973) A meditative, sultry, and entirely experimental horror film, writer/director Bill Gunn's Ganja stars Duane Jones (Night of the Living Dead) as Dr. Hess Green, who is attacked by his assistant with a knife belonging to a fictional ancient African tribe. He’s subsequently compelled to drink the blood of his assailant, an act that completes his vampiric transformation. Shortly thereafter, the assistant’s wife shows up looking for her husband and, even once she realizes what happened, begins a love affair with the doctor. Stylish and deliberately paced (a lot of vibes; a little plot), the movie has some smart, and very barbed, points to make about assimilation and religious hypocrisy. Spike Lee remade it in 2014 as Da Sweet Blood of Jesus but, with respect to Lee, you're much better off with the original. You can stream Ganja & Hess on Tubi and Kanopy or rent it from Prime Video. Ganja & Hess (1973) at Prime Video Learn More Learn More at Prime Video A Girl Walks Home Alone at Night (2014) Iranian-American writer/director Ana Lily Amirpour's A Girl Walks Home Alone at Night would make a list of cool vampire movies on its gorgeous black-and-white style alone: Gliding down the streets of an Iranian town in her chador, The Girl (Sheila Vand) cuts an iconic figure, even when we learn that her eerie way of movement has to do with a concealed skateboard. But Amirpour has a lot of fun with the idea that a girl traveling a deserted street in the dark might be a threat far more than a potential victim, and that a chador might be less a symbol of oppression than a means of staying concealed until the moment comes to strike. You can stream A Girl Walks Home Alone at Night on Kanopy or rent it from Apple TV+. A Girl Walks Home Alone at Night (2014) at Apple TV+ Learn More Learn More at Apple TV+ Dracula's Daughter (1936) Like father, like daughter in this direct sequel to the Bela Lugosi film, with Countess Marya Zaleska (Gloria Holden, in full Garbo mode) running afoul of Edward Van Sloan's Van Helsing from the original. The two vampires share in common not just an archnemesis, but an eye for the ladies, as well, with Countess Zaleska seducing first despondent Lili and then kidnapping Janet. What it lacks in style (not being nearly a match for Todd Browning's film in that regard) it makes up for in sheer horniness. Producers and censors were nervous about the lesbian of it all in the lead-up to the film's release, but also chose to hype it up in the marketing, especially with the very on-the-nose tagline: "Save the women of London from Dracula's Daughter!" You can rent Dracula's Daughter from Prime Video. Dracula's Daughter (1936) at Prime Video Learn More Learn More at Prime Video Bit (2019) Nicole Maines (Supergirl) stars here as Laurel, a trans teenage girl who's off to sunny LA following her transition. She's looking for a fresh start, and it's all going great until a club afterparty leads to lots of making out—and then some blood drinking. It’s not long before Laurel is offered the chance to join an all-female group of vampires dedicated to taking care of the many predatory men walking the city’s streets. You can stream Bit on Prime Video. Bit (2019) at Prime Video Learn More Learn More at Prime Video What We Do in the Shadows (2014) Starring, written, and directed by Jemaine Clement and Taika Waititi, Shadows precedes the equally great show of the same name with a similar premise: A houseful of old-world vampires (with old-school vampire trappings) looks desperately uncool when faced with the modern world. Many vampire narratives derive pathos from the idea of people trapped in a state of dull, unchanging immortality. Shadows takes that notion and mines it for solid laughs. You can rent What We Do in the Shadows from Prime Video. What We Do in the Shadows (2014) at Prime Video Learn More Learn More at Prime Video Daughters of Darkness (1971) Erotic bisexual vampire movies, believe it or not, had a moment in the 1970s; few (if any) were more successful, and more genuinely sexy, than Daughters of Darkness. The movie has a look and feel that borders on arthouse, with a grand and elegant style, as well as a willingness to go deeper with themes (western decadence, gendered power dynamics, etc.) that have been discussed in many an academic context in the decades since. If all that makes it sound like a bore, I'll repeat that it's still very much an erotic bisexual vampire movie. Delphine Seyrig plays Countess Elizabeth Báthory, who happens upon a newlywed couple honeymooning in a remote region, and immediately sets about seducing the wife away from human sexual and moral conformity. You can stream Daughters of Darkness on Tubi and Shudder or rent it from Prime Video. Daughters of Darkness (1971) at Prime Video Learn More Learn More at Prime Video The Lost Boys (1987) It might not be the gayest vampire movie (there's a lot of competition), but it may well be the sweatiest. Directed by out gay director Joel Schumacher (several years before he made the gayest superhero movie, Batman & Robin), The Lost Boys finds brothers Sam (Corey Haim) and Michael (Jason Patric) Emerson moving with their family to sunny, schvitzy Santa Carla, California—just in time for Michael to be seduced into joining a sexy vampire gang by its blonde leader David, played by Kiefer Sutherland. The vamps bait Michael with Star (Jamie Gertz), but his desire for the girl gives way to a fascination with David that leads him down a very bloody path. You can rent The Lost Boys from Prime Video. The Lost Boys (1987) at Prime Video Learn More Learn More at Prime Video Cronos (1992) Guillermo del Toro begins in this deeply unconventional vampire movie, the writer/director's debut. In the 16th century, an alchemist invents the titular device: a steampunk-ish scarab that conceals a creature inside. When it's rediscovered in the present by Jesús Gris (played by del Toro favorite Federico Luppi), it injects the man with an elixir that makes him (a bit) younger, and gives him a vampire's thirst for blood. Complicating matters, a dying oligarch who'd been on the trail of the Cronos isn't very happy that someone else gets the use of its powers, so he sends his relentless thug of a nephew (Ron Perlman) to hunt it down. The imagery is weird and endlessly inventive and, what's more, del Toro injects enough dark comedy around the edges so that the proceedings never bog down in self-seriousness. You can stream Cronos on Max and The Criterion Channel or rent it from Prime Video. Cronos (1992) at Prime Video Learn More Learn More at Prime Video Blade II (2002) A decade on from Cronos, Guillermo del Toro directed this rather brilliant action spectacle—a blood-soaked good time that blends its disparate elements into a propulsive thrill ride that knows exactly what kind of movie it is. Wesley Snipes is at his sassy best as the title's vampire/vampire hunter, while every set piece is bright and colorful, and the vampires are thoroughly monstrous. Practical effects and an artist's eye turn a comic-book action movie into an essential. You can rent Blade II on Prime Video. Blade II (2002) at Prime Video Learn More Learn More at Prime Video Nadja (1994) A fascinating novelty from director Michael Almereyda, Nadja is based (roughly) on André Breton's French surrealist manifesto of the same name; when financing for such an adaptation fell through (a near-instantaneous process), David Lynch (who also appears) stepped in and paid for the whole thing. To save money and achieve a particular look, Almereyda filmed vampire point-of-view shots on a Fisher-Price camera. All of which is to say that this is unlike other vampire movie. Two plot points kick off the film: First, the death of Count Dracula at the hands of his old nemesis, Van Helsing (Peter Fonda)—traumatizing his daughter (Elina Löwensohn); then, that same Nadja seeking out the daughter of Van Helsing to have sex with so that she can make her a thrall and carry out her goal of getting revenge on the Van Helsing family. It's an arty take on blood, gore, and horny lesbian vampires. You can stream Nadja on Prime Video. Nadja (1994) at Prime Video Learn More Learn More at Prime Video Thirst (2009) "What if we took Emile Zola’s non-supernatural 1868 novel Thérèse Raquin and made it into a vampire horror film?" said writer/director Park Chan-wook (Oldboy), apparently. The novel is about an orphan forced into a miserable marriage with a droopy hypochondriac who begins a torrid, ill-fated affair. Here, the orphan is Tae-ju (Kim Ok-vin), and the dashing man with whom she's having an affair is Sang-hyun (Song Kang-ho), a one-time Catholic priest who developed vampiric hungers following an experiment gone wrong. Wildly stylish, and fascinating in being surprisingly faithful to a book that has nothing at all to do with vampires. You can rent Thirst from Prime Video. Thirst (2009) at Prime Video Learn More Learn More at Prime Video Only Lovers Left Alive (2013) Is Only Lovers Left Alive a horror movie? A Gothic romance? A stoner-ish comedy? Um...yes? Indie writer/director Jim Jarmusch, who'd already put his distinctive stamp on other genres (like the idiosyncratic western Dead Man), turned his camera to horror. The movie follows a vampire couple, both alive for centuries but now living a world apart: Adam (Tom Hiddleston) is a sullen and moody musician living in a trashy Detroit apartment, while Eve (Tilda Swinton) lives a cheerier life in Tangier. She brings Adam along with her to give him a change of scenery and get him out of his gloomy rut, where she encourages him to get out to live a little, or whatever the undead equivalent is. You can rent Only Lovers Left Alive from Prime Video. Only Lovers Left Alive (2013) at Prime Video Learn More Learn More at Prime Video Interview With the Vampire (1994) It's great that the modern TV-series adaptation dispenses with the subtext and gets real as to the relationship between Lestat and Louis, but this 1994 adaptation walks impressively close to that edge. Vampires Lestat (Tom Cruise) and Louis (Brad Pitt) set up housekeeping in moody New Orleans, "adopting" daughter Claudia (Kirsten Dunst) and generally making a glorious mockery of the all-American family unit. You can rent Interview with the Vampire from Prime Video. Interview With the Vampire (1994) at Prime Video Learn More Learn More at Prime Video Martin (1977) George Romero manages to pack an extensive exploration of vampire lore into a film that may or may not even involve a vampire—it's not much a spoiler to offer that we're never entirely clear whether the vampirism of Martin (John Amplas) should be taken literally. From a family steeped in Eastern European folklore and myth, Martin develops a driving need to drink blood, a need that leads him to drug his victims and then slice into their veins with a razor. As usual, Romero has a lot to say, but here he's expertly playing on notions of vampires as sexual stand-ins (the impotent Martin role-plays the highly sexualized vampire attacks of earlier movies), and Romero also invites us to question our fascination with Martin even as he's engaged in repulsive acts. You can stream Martin on Tubi. Martin (1977) at Tubi Learn More Learn More at Tubi Let the Right One In (2008) One of the finest vampire movies, certainly of this century, is unexpectedly poignant, even as it doesn't skimp on the brutality. Bullied Oskar (Kåre Hedebrant) meets and befriends Eli (Lina Leandersson), a girl who appears to be about his age, but who is (you won't be surprised to learn) a vampire. Though very different, the two bond over their shared status as outsiders in the gloomy Swedish suburb of Blackeberg. You can stream Let the Right One In on Kanopy or rent it from Prime Video. Let the Right One In (2008) at Prime Video Learn More Learn More at Prime Video Stake Land (2010) Stake Land's vampires are neither sexy nor subtle. They're fully animal creatures that stalk an apocalyptic wasteland—cutting to the chase as to what a world with vampires might quickly come to look like. As a thriller, it's gory and effective, but a movie just about humans on the run from scary vampires would be less interesting than what we get here: As is so often the case, it's the humans who remain the scariest threats. Religious fanatics calling themselves The Brotherhood are convinced that vampire are God's will, a punishment for sins and, as a result, they're perfectly happy to help the hungry beasts feed. You can stream Stake Land on Tubi, Kanopy, Peacock, and Prime Video. Stake Land (2010) at Peacock Learn More Learn More at Peacock Blood Relatives (2022) We're talking about cool vampires here, but this comedy from director Noah Segan is about a terminally uncool vampire (played by Segan), who discovers that he's the father of a grown young woman (played by a great Victoria Moroles) who inherited his vampirism, in part, and wants to connect with her dad following the death of her mom. Mostly, she just has nowhere else to go. In and among the snappy dialogue and sharp characterizations, Segan has fun puncturing vampire lore in the story of a guy who initially has no interest in connecting with his daughter, but then has a change of heart. You can stream Blood Relatives on Shudder or rent it from Prime Video. Blood Relatives (2022) at Prime Video Learn More Learn More at Prime Video Dracula (1931) If it's hard to imagine a time when horror wasn't a powerhouse movie genre, that's because it's been that way since 1931, the year that Béla Lugosi's Count Dracula sashayed onto the screen. Tod Browning's film was a prestige adaptation of a classic novel, before it occurred to anyone to treat horror as B-level entertainment, so there's no corner-cutting in Dracula's impeccable style. Though it gets a bit stagey in its second half, the movie gives us some of horror's most indelible imagery—scenes and shots we're still referencing nearly a century later. Béla Lugosi's Dracula has been endlessly parodied and imitated, but all of that's only possible because he made one hell of an impression. You can rent Dracula from Prime Video. Dracula (1931) at Prime Video Learn More Learn More at Prime Video View the full article

Kohl’s has terminated its new CEO Ashley Buchanan after an investigation determined that he directed the retailer to engage in vendor transactions that involved undisclosed conflicts of interest. Kohl’s named Chairman Michael Bender as interim CEO, effective immediately. In connection with the appointment, Bender will step down as a member of the board’s audit, compensation and nominating and environmental, social and governance committee, according to the retailer’s regulatory filing. The news comes nearly four months after Buchanan, who had been previously the CEO of arts and crafts chain Michaels, took over the job on January 15. Buchanan’s appointment marks the third CEO for Kohl’s in three years as the department store struggles to reverse sluggish sales. Kohl’s said Thursday that Buchanan’s firing is unrelated to its performance, financial reporting, results of operations and did not involve any of its other employees. Kohl’s will conduct a search for a permanent CEO and said it will name a new chair in due course. The company couldn’t be immediately be reached for comment. Buchanan didn’t immediately return a message sent to his LinkedIn account. According to the Securities and Exchange Commission filing, Buchanan’s termination follows a probe conducted by outside counsel and overseen by the board’s audit committee. It found Buchanan had directed that Kohl’s conduct business with a vendor founded by an individual with whom Buchanan has a personal relationship on “highly unusual terms favorable to the vendor” and that he also caused Kohl’s to enter into a multimillion-dollar consulting agreement with the same individual who was a part of the consulting team. It also found that in neither case did Buchanan disclose this relationship as required under Kohl’s code of ethics. In connection with his termination and in accordance with the terms of his equity award agreements, Buchanan will forfeit all equity awards he received from the company, including the recruitment awards made as of January 15, according to the filing. Buchanan will also be required to reimburse Kohl’s for a pro rata portion of his signing incentive in the amount of $2.5 million, according to the documents. As a result of Buchanan’s termination, the board has determined to withdraw his nomination for election as a director of the company at the company’s annual shareholders’ meeting to be held on May 14. Buchanan had succeeded Tom Kingsbury, who stayed on as an adviser and is retaining his position on Kohl’s board until his retirement next month. Kingsbury served as Kohl’s interim CEO in December 2022 and was named its permanent leader in February 2023. The firing comes at a time when Kohl’s, which operates 1,600 stores across the country, is wrestling with sluggish sales. Its middle-income shoppers have pulled back on discretionary spending in the face of still-high prices for necessities. It’s also faced stiff competition from Walmart and Amazon, which have been improving their fashion offerings at affordable prices. And like other retailers, it is confronting uncertainty surrounding President Donald The President‘s expansive tariffs. On Thursday, Kohl’s offered a preliminary look at sales and profits for the current quarter that showed continued weakness, though the expected results are on track to beat Wall Street estimates. It said that it expects to report a decline in comparable sales—those coming from established physical stores and online channels—in the range of 4.3% to 4%, and a loss of 24 cents to 20 cents per share for the fiscal first quarter. Analysts expected earnings per share loss of 54 cents and a drop in comparable sales of 6.4%, according to FactSet. It expects to report final fiscal first-quarter results on May 29. Shares of the company, based in Menomonee Falls, Wisconsin, rose nearly 9% in late morning trading. —Michelle Chapman and Anne D’Innocenzio, AP business writers View the full article

US intelligence agency seeks to woo prospective assets with the offer of ‘a better life’View the full article

Hackers target luxury department store group after Marks and Spencer and the Co-op View the full article

From The President Tower to the Vatican: inside the tortuous dealmaking to link Ukraine’s resources to US supportView the full article

Google started placing ads inside chat conversations with some third-party AI assistants, according to a Bloomberg report, marking another step in its push to monetize AI-powered search alternatives. Driving the news. Google’s AdSense network — traditionally used for placing ads in search results and across websites — is now running ads within chatbot interactions, according to people familiar with the rollout. The company reportedly began testing the feature earlier this year with conversational AI search startups iAsk and Liner. A Google spokesperson confirmed that AdSense is available “for websites that want to show relevant ads in their conversational AI experiences.” Why we care. Google’s move brings targeted advertising into a new, fast-growing channel: AI chatbots. As users increasingly turn to conversational AI for answers instead of traditional search, placing ads in these interactions ensures brands stay visible where attention is shifting. This also opens up new, highly contextual ad formats that could deliver better engagement and insights than standard web placements. The big picture. AI chatbots like ChatGPT, Claude, and Perplexity are reshaping how people get answers online – a space historically dominated by Google Search. Google has responded by integrating its own large language models into products like Gemini and Search. Last year, it began including ads in its AI-generated overviews on Search results pages. Between the lines. By partnering with AI startups, Google can capture ad revenue regardless of who owns the search experience. Liner tailors a small number of ads to longer user queries, a model CEO Luke Jinu Kim likens to an “early version of Google search ads.” iAsk places ads below AI-generated answers before prompting follow-up questions. Yes, but. AI-driven platforms often see lower click rates – a challenge for advertising models built on volume and engagement. Still, research-heavy use cases like Liner’s may offer more targeted opportunities for brands. What to watch. Whether users find ads in AI chats useful or intrusive, and how Google balances monetization with user trust in an evolving AI search landscape. View the full article

For the second time in recent months, the Food and Drug Administration is bringing back some recently fired employees, including staffers who handle travel bookings for safety inspectors. More than 20 of the agency’s roughly 60 travel staff will be reinstated, according to two FDA staffers notified of the plan this week, who spoke on condition of anonymity to discuss confidential agency matters. Food scientists who test samples for bacteria and study potentially harmful chemicals also have been told they will get their jobs back, but have yet to receive any official confirmation. The same uncertainty hangs over employees who process agency records for release to lawyers, companies and journalists under the Freedom of Information Act. About 100 of those staffers were recently eliminated, according to an agency official with direct knowledge of the situation. But in recent days the FDA has missed multiple court-ordered deadlines to produce documents, which could result in hefty fines. That’s prompted plans to bring back a significant number of those staffers. The apparent reversals are the latest examples of the haphazard approach to agency cuts that have shrunk FDA’s workforce by an estimated 20%, or about 3,500 jobs, in addition to an unspecified number of retirements, voluntary buyouts and resignations. In February, the FDA laid off about 700 provisional employees, including food and medical device reviewers, only to rehire many of them within days after pushback from industry, Congress and other parties. The Department of Health and Human Services hasn’t detailed exactly which positions or programs were cut in the mass layoffs. FDA Commissioner Marty Makary has repeatedly said that no FDA scientists were fired as part of the reductions. But at least two dozen food scientists who worked in a San Francisco testing laboratory and a Chicago research center were let go in March. An HHS spokesperson suggested the apparent mix-up was due to “the fractured, outdated HR infrastructure we inherited from the Biden administration and are now actively overhauling.” The spokesperson did not respond to specific questions about which employees are being reinstated but said the administration will “streamline operations and fix the broken systems left to us.” About 15 scientists working in FDA’s Division of Food Processing Science and Technology in Chicago were told last week they be will reinstated, according to a staffer who spoke on condition of anonymity to discuss confidential agency matters. But a week later there has been no written confirmation and the scientists have not returned to the office. The group’s research includes studying ways to prevent harmful bacteria from growing on produce and preventing the spread of microplastics and other particles from food packaging. “I hope Commissioner Makary continues to assess these ill-informed cuts and works to bring back impacted employees expeditiously,” said Susan Mayne of Yale University, the FDA’s former food director. “His legacy as commissioner is on the line.” With more than 15,000 employees remaining across various U.S. and foreign offices, the FDA’s core responsibilities are reviewing new drugs, medical products and food ingredients as well as inspecting thousands of factories. Makary has said no inspectors or medical reviewers were fired as part of the recent reductions. But current and former FDA officials note that those frontline employees are often supported by teams of administrative staff. FDA inspectors, for example, have long relied on travel bookers to coordinate trips to India and other countries that often involve visa permissions, security measures, ground transportation, tech support, translation services and other logistics. Inspectors can spend up to half the year traveling, a grueling workload that makes recruiting and retaining staff a challenge. For a brief period last month, inspectors were told they would be booking their own travel. The FDA set up a hotline to assist with making the arrangements. Then, agency leaders developed a plan to hire an outside contractor to perform the work. On Monday, staffers were informed that about a third of the fired staff who performed the work would be returning. The Associated Press Health and Science Department receives support from the Howard Hughes Medical Institute’s Science and Educational Media Group. The AP is solely responsible for all content. —Matthew Perrone, AP health writer AP reporter JoNel Aleccia contributed to this report. View the full article

... and lots of it. By CPA Trendlines Research Go PRO for members-only access to more CPA Trendlines Research. View the full article

Scams are a dime a dozen these days—so much so, you're pretty safe ignoring most of the messages you receive from unknown senders. But every now and then, an uncommon message isn't a scam, spam, or malicious in intent; in fact, sometimes, it's an official message from none other than Apple, warning you that you're being spied on. As much as this situation sounds like classic spam, it's very much not: Apple actually did send alerts to users on Wednesday warning them they might be targeted by "mercenary spyware attacks." Two of the users Apple alerted were Ciro Pellegrino, an Italian journalist for Fanpage, and Eva Vlaardingerbroek, a Dutch right-wing activist. You can see an excerpt of the alert below: “ “Apple detected that you are being targeted by a mercenary spyware attack that is trying to remotely compromise the iPhone associated with your Apple ID -xxx-...This attack is likely targeting you specifically because of who you are or what you do. Although it’s never possible to achieve absolute certainty when detecting such attacks, Apple has high confidence in this warning — please take it seriously.” ” This isn't the first time Apple has sent users this type of warning. In fact, the company sent a round of warnings to users back in July, complete with the same message as noted above. They did the same in April of 2024, as well. According to an Apple support document on the subject, the company has sent alerts multiple times each year since 2021, now to over 150 countries in total. The specifics of the attempted attacks are limited, since Apple doesn't want to reveal how it is able to detect the risks to users. While that's understandable, it means users are left knowing they're likely the target of a spyware campaign, but without knowing from who or where. To be clear, the vast majority of users receiving these alerts are not your everyday iPhone customers. Most of the users Apple believes are being targeted by spyware work in high-profile positions that attract attacks from state actors. Think politicians, journalists, activists, diplomats, etc.—people who expose secrets or have power that other people in power would like to stop. Bad actors spend millions of dollars to target these users in elaborate spyware campaigns, attempting to install malware on their devices in an effort to spy on location, data, and activity. Perhaps the most notable spyware of this kind is Pegasus, which was created by the Israeli-based NSO Group to spy on a host of high-profile targets the government sees as "threats." What makes spyware like Pegasus particularly dangerous is it requires no action on the part of the target: Bad actors can infect iPhones and Androids with Pegasus silently in what's known as a "zero-click" attack, and track just about everything the target does on their smartphone. What to do if you receive this alertIf you do receive one of these alerts, the first step is to ensure it's real. The company will send alerts via iMessage and email: iMessage alerts are sent from [threat-notifications@apple.com], while email alerts are from [threat-notifications@email.apple.com]. Before April of this year, Apple sent both from the @apple.com domain. Even if you confirm the address, sign into appleid.apple.com and look for the threat notification at the top of the page. If Apple really sent you one, you'll see it here. Otherwise, assume the message is fake. Apple says its alert will never ask you to click a link, open a file, or install an app or profile, so if the "alert" asks you to do this, ignore it. These are classic tactics to trick users into installing the very malware a true alert would try to protect you from. The company also recommends you reach out to the non-profit Access Now and use their Digital Security Helpline. While they won't be able to offer you specific advice about your situation, they will be able to walk you through general steps to secure yourself. Whether you receive an alert, or your work involves one of these at-risk positions, Apple strongly encourages you to enable Lockdown Mode on your Apple devices. Lockdown Mode restricts many of the basic functions of your Apple devices, to plug potential holes bad actors can exploit to compromise those devices. This includes blocking things like message attachment types, web technologies in Safari, and incoming FaceTime calls. The feature also removes your location from shared photos, and stops configuration profiles from being installed on your device. As Lockdown Mode limits the features of your iPhone or Mac, it's not something that most people should use on a daily basis. However, for those who may be targeted by bad actors, it can be a great line of defense. You can follow our guide here to enable Lockdown Mode on your Apple devices to protect yourself. View the full article

... and lots of it. By CPA Trendlines Research Go PRO for members-only access to more CPA Trendlines Research. View the full article

Some good news for Publishers Clearing House (PCH) customers: The Federal Trade Commission (FTC) said Wednesday that the sweepstakes company is paying for refund checks to 281,724 customers who ordered a product after receiving and clicking on an email, which included “deceptive and unfair” practices, according to the FTC’s allegations. Here’s what you need to know. What happened? Publishers Clearing House, which has been known for decades for its sweepstakes deals and big checks, agreed to pay a total of about $18.5 million in refunds and make substantial changes to its e-commerce operations. “While we disagreed with the FTC’s assertions at the time, we were glad to have resolved the matter and move forward continuing to do what we do best—provide consumers fun entertainment and games powered by our famous chance to win,” Christopher Irving, vice president of consumer and legal affairs at PCH, told CBS News. Irving added that the current refunds are based on the FTC’s complaint and settlement from two years ago. Among other things, the 2023 lawsuit alleges PCH misled customers into thinking they had to make a purchase to enter the sweepstakes or to increase their chance of winning, and that the email subject lines were deceptively worded. The FTC also charged that the company added surprise shipping and handling fees to the costs of products, misrepresented that ordering is “risk free,” used deceptive emails as part of its marketing campaign, and misrepresented its policies on selling users’ personal data to third parties prior to January 2019. Here are the three main charges of the FTC’s complaint: PCH targeted older and lower-income consumers, deceiving them into thinking either that consumers could not enter into sweepstakes without purchasing a product, or that their chances of winning would be increased by purchasing products. PCH misled consumers by sending emails with deceptive subject lines that led consumers to believe the email was related to official documents, such as tax forms. PCH added deceptive shipping and handling fees and misrepresented that ordering was “risk-free,” even though consumers who wanted refunds had to return products at their own expense. What to know about getting a Publishers Clearing House refund check The FTC said recipients who receive a refund should cash their checks within 90 days, as indicated on the check. If you received a refund and have additional questions, contact Publishers Clearing House at 888-516-0774 or by email here. Consumers who have questions about the refund process can also visit the FTC website to view the frequently asked questions. The FTC does not require people to pay money or provide account information to get a refund. For some context, in 2024, the commission’s actions led to more than $338 million in refunds to consumers across the United States. View the full article

The The President administration has framed tariffs as a necessary tool for bringing more jobs to the U.S. and reviving the manufacturing sector. But many economists have warned that widespread job creation is unlikely, given the cost to companies—and in the meantime, The President’s substantial tariffs will drive up prices for both consumers and businesses, likely forcing them to cut costs through layoffs. Many frontline workers have already expressed concerns about the effect that the tariffs could have on their job security. In a survey by workforce management software company UKG, over half of the 5,000 frontline workers surveyed—defined as people who do shift work or are paid hourly—said they believed they could be laid off, while 74% expect that the tariffs will affect their earnings potential. Gen Z workers were the most likely to be concerned about layoffs, but the majority of workers described feeling nervous, stressed, or angry about the impact of tariffs on their jobs. Though the tariffs have already shaken up financial markets, the vast majority of workers (77%) believe that The President’s trade policies will harm smaller businesses more than Wall Street firms. According to the survey, tariffs are also driving changes in how workers are showing up on the job. Over 70% of respondents said their workplace behavior had changed in some capacity: Many of them claimed to be working harder to “prove their value,” while others were picking up additional shifts. Nearly half of workers were striving to increase their savings. About two-thirds of those surveyed said they expected the tariffs would likely limit their future job prospects. Workers have reason to be worried. President The President’s trade policies already seem to be impacting the workforce: The automaker Stellantis has trimmed headcount by about 900 across several manufacturing plants anticipating the impact of tariffs, while Volvo is cutting up to 800 jobs. Just this week, UPS announced that it would slash 20,000 jobs within the year to reduce costs, citing “macroeconomic uncertainty” and also noting the high likelihood of decreased shipping volume from China due to the tariffs. (Another major factor is that UPS is significantly cutting back on deliveries for Amazon.) Agricultural exporters are feeling the financial effects of the tariffs and have turned to layoffs, according to a CNBC report. Some experts have said that the tariffs might eventually create more manufacturing jobs stateside, and a number of major companies have already said they are expanding their manufacturing footprint in the U.S. But a Goldman Sachs analysis found that the tariffs could also lead to hundreds of thousands of job losses across the workforce—something that many workers clearly seem to anticipate. View the full article

Organizations look structured and logical from the outside—boxes and lines, reporting relationships, KPIs, and performance frameworks. But walk into any real meeting, and you’ll sense it: side glances, shifting energy, people going silent when one voice enters the room, unexplained resistance to change, and power dynamics no slide deck could predict. That’s not just dysfunction. That’s the system speaking, and most leaders aren’t listening. That is why we need something called systemic intelligence. Systemic intelligence is the capacity to sense and respond to the invisible forces shaping an organization’s behavior, culture, and outcomes. It’s not about titles or tactics. It’s about understanding: The unspoken agreements that guide behavior The loyalties people carry—to past leaders, ideas, or roles The emotional undercurrents in teams and across departments The patterns of inclusion and exclusion that shape decision-making The stories that are being told, and the ones that aren’t If emotional intelligence helps you understand individuals, systemic intelligence enables you to understand relationships, fields, and patterns. It’s what allows a leader to walk into a room and feel the temperature, not just the metrics, but the mood of an organization. Why This Matters More Than Ever The modern workplace is in flux. Hybrid work, generational shifts, AI transformation, and rising emotional exhaustion reveal how fragile many organizational systems are. And yet, most leadership development still focuses on logic, linearity, and surface-level skills. Here’s the reality: 70% of transformation efforts fail, primarily due to hidden dynamics—cultural resistance, misalignment, and lack of trust. Furthermore, only 27% of employees believe their company’s values align with how work actually gets done. Most strategies fail not because they are wrong but because they are disconnected from the reality of the system they are trying to move. If leaders don’t learn to see the system, they will be ruled by forces they don’t understand. A Moment That Changed Everything I once worked with a leadership team navigating the aftermath of a merger. They had a new vision, a reorg plan, and a glossy set of PowerPoint decks. But something was stuck. Meetings were tense. Morale was low. Alignment felt forced. So, we paused the strategy session and held a story circle. One leader finally voiced what everyone else had been feeling: “I still feel loyalty to our former CEO. We never really said goodbye. And it feels like we’re not allowed to grieve the culture we lost.” In that moment, something shifted. What emerged wasn’t just emotion; it was clarity. The energy in the room softened, and trust began to rebuild. The team could finally move forward—not by pushing harder, but by acknowledging what had been in the system all along. What you don’t name, you can’t shift. The S.E.E.N. Framework for Systemic Intelligence Systemic intelligence isn’t about having special powers. It’s about cultivating a new kind of leadership presence that’s attuned to what’s happening beneath the surface. You don’t develop this awareness by accident. You create it by practicing small but powerful shifts in observing, listening, and engaging with your organization as a living, breathing system. To help leaders begin, I use a simple guide: “S.E.E.N.” It’s a reminder that before you can shape a system, you must first learn to see it. S – Sense the Field. Slow down. Listen beyond the words. What’s present, but unspoken? What’s the emotional temperature? Before jumping into action, ask your team: “What’s the mood in the room right now?” Then sit with the silence. E – Explore Hidden Loyalties. People don’t just commit to goals—they commit to identities, past leaders, and unspoken rules. What loyalties are operating beneath the surface? For example, a team resistant to innovation may not fear change—they may be protecting the legacy of a beloved product or person. E – Examine the Energy Flow. Where is energy stuck? Who gets centered, and who gets sidelined? Where does attention naturally go? Where does it get blocked? Map informal influence—not just reporting lines. Who really holds trust in the system? N – Name What Needs to Be Acknowledged. Often, healing doesn’t come from solving—it comes from witnessing. What grief, transition, or injustice needs to be seen and honored? What if your next strategic move began with a ritual of acknowledgment, not another set of objectives? How to Start Seeing the System You don’t need to become a therapist. You just need to become more attuned to the emotional undercurrents, unspoken dynamics, and patterns shaping your team. Here are a few ways to begin: Host Campfire Conversations. Create spaces where stories—not just updates—can be shared. Start with: “Tell us about a moment that shaped your connection to this organization.” Bring in Outside Eyes. Artists, facilitators, systemic coaches, or organizational psychologists can help visualize dynamics your team may be too close to see. Use Visual Mapping. Ask: “What’s the formal structure? What’s the informal one? Who’s at the center of decisions, and who’s on the margins?” Slow the Agenda. Build in white space. Let emotion, silence, or discomfort have a seat at the table. Intelligence lives in the spaces we’re often too quick to fill. Most leaders try to fix what they can see. But true leadership begins by learning to sense what you can’t. Strategy is important, and structure is necessary, but without systemic intelligence, even the best plans will stall. Because what’s unacknowledged gets acted out, and what’s seen can finally start to shift. So, the next time your team feels stuck, ask yourself: “What’s really going on here? What’s in the system that no one is naming?” That question might be your most strategic move yet. View the full article

Pricing on the 30-year fixed rate mortgage retreated this week as investors digested some economic news, including a GDP contraction in the first quarter. 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. Ever since Google acquired Fitbit, it has managed to improve its Pixel Watch series. The Google Pixel Watch 2 is a great mix of the Fitbit health-tracking features and Google's tech-heavy smartwatch and is a capable smartwatch option for Android users even in 2025. After Google announced the Pixel Watch 3 along with the Pixel 9 and Pixel Buds Pro 2 last summer, the predecessor received a hefty discount. Right now, the LTE Google Pixel Watch 2 is $199.99 (originally $399.99 at release), matching the lowest price it has been on Amazon since its October 2023 release, according to price-checking tools. Google Pixel Watch 2 Heart Rate Tracking, Stress Management, Safety Features - Obsidian Active Band - LTE $199.99 at Amazon /images/amazon-prime.svg $299.99 Save $100.00 Get Deal Get Deal $199.99 at Amazon /images/amazon-prime.svg $299.99 Save $100.00 According to PCMag's "excellent" review, the Pixel Watch 2 offers 33 hours of battery time, a Qualcomm 5100 processor, a bright AMOLED screen, and a multitude of features, including an accelerometer, GPS, a blood oxygen monitor, a temperature sensor, and others. The smartwatch is rated IP68 for dust and water resistance (you can submerge it up to 164 feet underwater). Its main competitor is the Samsung Galaxy Watch 6, which costs $234.99 (originally $299.99) for the Bluetooth version and offers a brighter display but a shorter battery life (22 hours according to PCMag). You can take calls with this LTE version, and you won't need to be within Bluetooth or wifi distance of your phone to use all of its features. If you take it with you on a run, you'll be able to leave your phone at home and you can still listen to your music, take texts, or calls. One of the biggest downsides of the Pixel Watch 2 is that it comes in a single size; a 41-millimeter screen, which is the same size as the small Apple Watch Series 9. Also, some of the Fitbit health features will require a premium subscription, which is $9.99 per month. If you like a comfortable watch with a battery that will last you more than a day and health-focused features like skin temperature and stress sensors, then the Pixel Watch 2 is a good option. View the full article

OpenAI has launched its latest models, o3 and o4-mini, described as the company’s smartest and most capable releases to date. The new models are now available to ChatGPT users with full access to tools like web browsing, file analysis with Python, visual input reasoning, and image generation. OpenAI says o3 is its most powerful reasoning model yet, delivering state-of-the-art performance in coding, math, science, and visual tasks. The model significantly reduces major errors compared to its predecessor, OpenAI o1, and shows notable improvements in programming, business consulting, and creative ideation. Meanwhile, o4-mini offers fast, cost-efficient reasoning. Optimized for high-volume use, o4-mini performs strongly in math, coding, and visual tasks, outperforming its predecessor, o3-mini, in benchmarks like AIME 2025 and data science evaluations. Both new models also demonstrate improved instruction-following capabilities and more personalized responses by referencing past conversations. For the first time, OpenAI’s models can think with images, integrating visual content into their reasoning processes. The models can analyze photos, charts, and low-quality visuals, using tools like rotation and zooming to assist problem-solving. OpenAI trained both models to not only use tools but also reason about when and how to deploy them. This allows the models to complete complex tasks independently, such as conducting web searches, writing code, generating graphs, and delivering full analyses—all within about a minute. Safety improvements accompanied these advances. OpenAI rebuilt its safety training datasets and deployed a reasoning LLM monitor to better detect dangerous prompts, reporting strong performance on internal refusal benchmarks. Evaluations confirmed that both o3 and o4-mini remain below OpenAI’s “High” risk thresholds across biological, cybersecurity, and self-improvement categories. Both models are now available to ChatGPT Plus, Pro, and Team users, replacing previous versions. Enterprise and Edu users will gain access within a week. Developers can access o3 and o4-mini through OpenAI’s Chat Completions API and the newly introduced Responses API. OpenAI also launched Codex CLI, a lightweight, open-source coding agent designed for direct terminal use, and announced a $1 million grant program to support Codex CLI projects, offering grants in $25,000 increments via API credits. OpenAI says its upcoming plans involve merging the strengths of the o-series and GPT-series models to create tools that combine natural conversation with advanced, proactive reasoning. Image: OpenAI This article, "OpenAI Introduces o3 and o4-mini Models with Enhanced Reasoning and Full Tool Access" was first published on Small Business Trends View the full article

OpenAI has launched its latest models, o3 and o4-mini, described as the company’s smartest and most capable releases to date. The new models are now available to ChatGPT users with full access to tools like web browsing, file analysis with Python, visual input reasoning, and image generation. OpenAI says o3 is its most powerful reasoning model yet, delivering state-of-the-art performance in coding, math, science, and visual tasks. The model significantly reduces major errors compared to its predecessor, OpenAI o1, and shows notable improvements in programming, business consulting, and creative ideation. Meanwhile, o4-mini offers fast, cost-efficient reasoning. Optimized for high-volume use, o4-mini performs strongly in math, coding, and visual tasks, outperforming its predecessor, o3-mini, in benchmarks like AIME 2025 and data science evaluations. Both new models also demonstrate improved instruction-following capabilities and more personalized responses by referencing past conversations. For the first time, OpenAI’s models can think with images, integrating visual content into their reasoning processes. The models can analyze photos, charts, and low-quality visuals, using tools like rotation and zooming to assist problem-solving. OpenAI trained both models to not only use tools but also reason about when and how to deploy them. This allows the models to complete complex tasks independently, such as conducting web searches, writing code, generating graphs, and delivering full analyses—all within about a minute. Safety improvements accompanied these advances. OpenAI rebuilt its safety training datasets and deployed a reasoning LLM monitor to better detect dangerous prompts, reporting strong performance on internal refusal benchmarks. Evaluations confirmed that both o3 and o4-mini remain below OpenAI’s “High” risk thresholds across biological, cybersecurity, and self-improvement categories. Both models are now available to ChatGPT Plus, Pro, and Team users, replacing previous versions. Enterprise and Edu users will gain access within a week. Developers can access o3 and o4-mini through OpenAI’s Chat Completions API and the newly introduced Responses API. OpenAI also launched Codex CLI, a lightweight, open-source coding agent designed for direct terminal use, and announced a $1 million grant program to support Codex CLI projects, offering grants in $25,000 increments via API credits. OpenAI says its upcoming plans involve merging the strengths of the o-series and GPT-series models to create tools that combine natural conversation with advanced, proactive reasoning. Image: OpenAI This article, "OpenAI Introduces o3 and o4-mini Models with Enhanced Reasoning and Full Tool Access" was first published on Small Business Trends View the full article

Critics warn costs could spiral into billionsView the full article

Google AI Mode is now available to all US users without a waitlist, offers new visual cards, shopping integration, and more. The post Google AI Mode Exits Waitlist, Now Available To All US Users appeared first on Search Engine Journal. View the full article

Big news for Peloton fans: As of this week, you can watch plain old YouTube on your Bike, Bike+, Tread, or Row. Now, you may be saying, "YouTube TV was already an option in my Peloton's Entertainment tab. What's the big deal?" The big deal, my friend, is that YouTube TV is a subscription television service and YouTube is, well, YouTube. This new integration allows you to watch standard videos on the platform, the same as you would on your phone or computer. How to access YouTube on the PelotonYou access YouTube the same way you access other non-class features, like scenic rides or my beloved Lanebreak game: through a special tab in the menu on the bottom of your screen. Specifically, it's in the Entertainment tab. (If you hit that tab and don't see it, don't worry. Sometimes, Peloton's software updates roll out in waves. You'll have it eventually.) Just tap YouTube and you'll be taken to a landing page that initially shows you Peloton's channel on the video-sharing platform. You can log into your YouTube profile by signing in with your Google Account or just use it as a guest. At the top of the screen, you'll be prompted to start a workout if you'd like your resistance, cadence, and output to be tracked and monitored while you ride and watch. Why I love this as a Peloton optionOver the years, I've kind of fallen out of love with watching shows or movies while I do my cardio. I just feel like I don't move with enough intensity when I'm that distracted, which is why I prefer classes or the Lanebreak game. That said, I know that lots of people prefer or even need to be distracted to get their cardio in and if the option to watch YouTube gets more people on the bikes or treadmills, I am all for it. I tested the feature out this morning, first signing into my Google account and then searching for what I wanted to watch. There are a lot of spin classes available on YouTube and I have friends who watch those while they ride non-Peloton bikes at home. As I've explained in my Peloton Bike review, the classes that are designed for your Peloton and available through the associated app are always your best option, since they're made with your specific equipment in mind and track how well you're doing with that equipment, but if you ever feel like watching a different class, this is now an option. Personally, I opted to watch something I knew would get me hyped up, so I put on Lady Gaga's cinematic masterpiece, 2011's "Marry the Night" video. I play this song in the spin classes I teach all the time, but the video is something extra special to me. It worked flawlessly. There was no lag, the image was crisp and clear, and I could see my cadence and resistance at the bottom of the screen the whole time, just like I can when I'm taking a proper Peloton class. Being able to watch one of my favorite music videos actually did push me pretty hard. It wasn't until my end-of-workout review that I realized how fast or hard I'd been going or how many calories I torched. The video on my screen really sucked me in. What to keep in mind if you're watching YouTube on your PelotonThis is an awesome option if you want to do something besides a class, Lanebreak, or a scenic ride, but bear in mind that you won't have any cues from pros while you do it unless you're watching a YouTube spin class. I hesitate to recommend this if you're super new, but I say this as an instructor who is worried about your safety. Take a few classes first, if you haven't already, so you have an understanding of proper form and technique. I also know a lot of people get sucked into heavy YouTube holes. It's awesome to ride for a long time, but be careful to set some limits or keep track of how long you've been at it. Hydrate frequently and take some breaks. This isn't the same as falling into a YouTube binge on your couch; you could over-exert yourself without realizing it. View the full article

Welcome to AI Decoded, Fast Company’s weekly newsletter that breaks down the most important news in the world of AI. You can sign up to receive this newsletter every week here. The President harms AI progress by warring with universities Donald The President has done a lot to antagonize universities in his first 100 days. He cut off federal research funding to institutions like Princeton, Columbia, and Harvard, citing their alleged tolerance of antisemitism on campus. He also threatened the authority of college accreditation bodies that require schools to maintain diversity, equity, and inclusion programs. But these actions directly undermine the administration’s stated goals of strengthening the U.S. military and helping American tech companies maintain their narrow lead over China in AI research. Since World War II, the U.S. government has maintained a deep and productive relationship with universities. Under the leadership of Vannevar Bush, director of the Office of Scientific Research and Development, the government channeled significant research funding into university labs. In return, it received breakthroughs like radar and nuclear technology. Over the decades, university researchers have continued to contribute critical innovations used in defense and intelligence, including GPS and the foundational technologies of the internet. Today, the government increasingly relies on the commercial sector—including major contractors like Boeing and General Dynamics, and newer firms like Palantir and Anduril—for defense innovation. Yet universities remain essential. Much of the most advanced AI research still originates from academic computer science departments, many of which are powered by international students studying at institutions like MIT, Stanford, and Berkeley. These students often go on to found companies based on research initiated in academia. Whether they choose to build their businesses in the U.S. or return to their home countries depends, in part, on whether they feel welcome. When international students see research funding threatened or videos of PhD students being arrested by ICE, staying in the U.S. becomes a less appealing option. In a recent conflict with Harvard, the Department of Homeland Security even demanded information on the university’s foreign students and threatened to revoke its eligibility to host them. In response, over 200 university and college presidents have condemned the administration’s actions and are exploring ways to resist further federal overreach. Rather than discouraging international researchers and students, the U.S. should be sending a clear signal: that it remains a safe, supportive, and dynamic environment for AI talent to study, innovate, and launch the next generation of transformative companies. The best AI agents may be powered by teams of AI models working together During the first phase of the AI boom, labs achieved big intelligence gains by pretraining their models on ever-larger data sets and using more computing power. While AI companies are still improving on the art and science of pretraining, the intelligence gains are becoming increasingly expensive. A big part of the research community has shifted its focus to finding the best ways to train models to “think on their feet,” or to reason over the best routes to a responsive and accurate answer at “inference time” just after a user enters a question or problem. This research has already led to a new generation of “thinking” models such as OpenAI’s o3 models, Google’s Gemini 2.0, and Anthropic’s Claude 3.7 Sonnet. Researchers teach such models to think by giving them a multistep problem and offering them a reward (usually just a bit of code that means “good”) for finding their way to a satisfactory answer. It’s certainly possible to build an inference system that makes numerous calls to a single large frontier AI model, collecting all the questions and answers in a “context window” as works toward an answer. But new research from Berkeley’s AI research lab shows this monolithic “one model to rule them all” approach is not always the best way of building an efficient and effective inference system. A “compound AI system” of multiple models, knowledge bases, and other tools working together can yield more relevant and accurate outputs at far lower costs. Importantly, such a “pipeline” of AI resources can be a powerful backend for AI agents capable of calling on tools and working autonomously, says Jared Quincy Davis of the AI cloud company Foundry. Foundry builds software that enables it to provide GPU compute at low cost for AI developers. Davis has led an effort to create an open-source framework that lets AI practitioners build just the right pipeline, with just the right resources, for the application they have in mind. The framework, called Ember, was created with help from researchers at Databricks, Google, IBM, NVIDIA, Microsoft, Anyscale, Stanford, UC Berkeley, and MIT. Davis says it’s possible to build a compound system that can make calls to a number of today’s state-of-the-art AI models (via APIs) such as ones from Google, OpenAI, Anthropic, and others. Large frontier models often stand above other large models in certain skill areas (Anthropic’s Claude is especially good at writing and analyzing text), so it’s possible to build a pipeline that calls on models according to their unique strengths. This is a very different way of looking at AI computing, compared to the narrative of just a couple years ago that said one model would be better than all others at practically everything. Now, numerous models compete for the state-of-the-art at various tasks, while other smaller models specialize in completing tasks at lower cost, and the overall cost of getting an answer from an AI model has gone way down over the past couple of years. Congress actually passes a tech bill Congress has failed to pass any broad-based regulation to protect user and data privacy on social networks. It has, however, managed to pass laws to prohibit specific and particularly dangerous social media content such as child sex trafficking, and now nonconsensual intimate images (or NCII). NCII refers to the practice of posting sexual images or videos of real people online without their consent (often as an act of revenge or an attempt to extort), including explicit images generated using AI tools. The bill, called the Take It Down Act, which unanimously passed the Senate in February and the House on Monday, makes it a federal crime to post NCII and requires that online platforms remove the content within 48 hours of a complaint. Affected “public-facing” online platforms will have a year after the law passes to set up a system for receiving and acting on complaints. The president is expected to sign the bill into law. Even though the bill’s intent earned widespread support, its legal reach disturbed some free expression advocates. The Electronic Frontier Foundation worries that the bill’s language is overbroad and that it could be used as a tool for censorship. These worries were compounded by the fact that the new law will be enforced by the Federal Trade Commission, which is now led by The President loyalists. More AI coverage from Fast Company: Duolingo doubles its language offerings with AI-built courses In his first 100 days, The President’s tariffs are already threatening the AI boom Microsoft thinks AI colleagues are coming soon Marc Lore wants AI to feed you—and make you healthier Want exclusive reporting and trend analysis on technology, business innovation, future of work, and design? Sign up for Fast Company Premium. View the full article

Google has a number of announcements around Al Mode, including dropping the waitlist, adding products and place cards to the answers, a new history panel, and that they are doing a limited test without having to opt in to Al Mode within Search Labs. Testing Al Mode in the wild. Google is now testing Al Mode within Google Search without you having to opt in to Al Mode within Search Labs. That means that you might see the tab for Al Mode without specifically opting into the feature. Soufi Esmailzadeh, Director, Product Management, at Google Search wrote, “Because our power users are finding it so helpful, we’re starting a limited test outside of Labs. In the coming weeks, a small percentage of people in the U.S. will see the Al Mode tab in Search, and we’ll continue to incorporate feedback into the experience. “ No waitlist. If you are eager to try out Al Mode and you have not yet been accepted into the waitlist, well, that waitlist is going away. Google will automatically approve anyone in the U.S. who is 18 years or older, if they opt into Al Mode. There is no longer a waitlist and you will get immediate access to the Al Mode tab. Places and products cards. Google also added places (local businesses) and product cards to AI Mode. This includes information such as local images, ratings, reviews, store hours, real-time prices, and availability. This pulls from both Google’s Shopping Graph and Google Business Profiles. Google wrote, “For local spots, like restaurants, salons and stores, you can quickly see info like ratings, reviews, and opening hours, and if you’re looking for a product, you’ll see shoppable options with real-time prices (including the latest promotions), images, shipping details and local inventory.” Here is what it looks like: History panel. Google also added a new history panel, so you can see your past conversations and queries. This allows you to quickly pick up from where you left off, directly in AI Mode. You can click the new left-side panel to get to your past searches. Google will topics, each will include the infomation that AI Mode already found for you, and you can ask follow-up questions or take your next steps. Why we care. Google and other search engines and AI companies, are looking for the killer AI search engines features that will gain them market dominance in this AI search race. AI Mode may be that interface for Google (or maybe not). That being said, Google is investing a lot into AI Mode; adding features and rolling it out more widely. Also, for Google to test this in the wild, it will give Google the information it needs to see how searchers really like AI Mode and then adapt it to make it better. It is exciting times in this AI Search space. View the full article