The number of women who hold executive reins in corporate Britain has fallen for the first time in eight years in an “unacceptable” reversal that could delay gender parity for another five generations, a consultancy said on Thursday.

The Pipeline, which analyses gender diversity at senior corporate levels, said the average proportion of women on executive committees at Britain’s 350 biggest, listed companies slipped to 32% in 2024 from a revised 33% last year.

The drop appears small. But Geeta Nargund, the group’s chair, told Reuters it reflected a sharp disconnect with male numbers and urged business leaders and headhunters to use the vast pool of female talent to fill decision-making roles, fix cultures and ensure women can thrive in the workplace.

“It is unacceptable that gender representation in business leadership is moving backwards in 2024 …,” she said.

“Organisations which are performing the best in terms of gender parity are 22% more likely to have improved profits … and so fair representation is not just a ‘nice to have’ or a tick-box exercise – it is a business imperative.”

Women still hold only 9% of chief executive (CEO) roles on FTSE 350 companies and 18% of top finance (CFO) positions – although they account for over 44% of chartered accountant roles, The Pipeline’s 2024 Women Count report showed.

Only 19% of women hold commercial boardroom jobs—roles that are accountable for a company’s profit and loss and can lead to jobs as a CEO and CFO—according to the data, sourced from research group BoardEx. That number is down from 20% in 2023.

The latest report comes after former Conservative Prime Minister Rishi Sunak‘s government stated last year that Britain was a leader for female representation because women held 40.2% of board positions on London’s top 350 listed companies – ahead of a 2025 deadline.

Britain’s Financial Conduct Authority (FCA) also requires listed companies to appoint women to at least 40% of board positions or explain why they have missed that target.

Board positions, however, include non-executive roles that lack operational power and decision-making functions.

—Kirstin Ridley, Reuters

MrBeast’s YouTube growth has hit a rough patch. 

On Saturday, MrBeast uploaded a new video. “7 Days Exploring An Underground City” amassed 56.5 million views within four days. For any other YouTuber, those numbers would be seismic. But, for YouTube’s most subscribed-to creator, the numbers are dramatically lower than his averages—and could be a sign of broader decline. 

“7 Days Exploring An Underground City” had all the trappings of a successful MrBeast video: A clickable premise, excited scream-speaking, and a gaggle of other YouTubers (Logan Paul, Kai Cenat) to help promote it. But the four-day view count of 56.5 million views is dramatically low for the channel. A MrBeast video hasn’t had that slow of viewer growth since May 2023. 

It’s been a slow backslide for MrBeast’s views in the past few months. Consider his last five videos: The earliest, published August 3, got 119 million views in four days. The following two videos earned 78 million views each in the same time frame. After that, his upload earned 67 million views. And now, the channel has hit a yearlong low of 56.5 million. 

MrBeast’s subscriber growth has also been down recently. In September, the YouTuber added five million subscribers to his channel. That’s the lowest it’s been in the past year, and far from the June peak of 27 million monthly added. It’s hard to imagine anyone overtaking MrBeast on this front; he remains 44 million ahead of the second-most subscribed-to channel, T-Series. 

MrBeast has the advantage of being a legacy channel, having grown his subscriber base since 2012. But his competitors are catching up. This past week, which includes the first two days of the new video’s release, MrBeast was only 7th in terms of views. T-Series, Toys and Colors, Sierra & Rhia FAM, and more were far ahead of him. 

MrBeast, whose real name is Jimmy Donaldson, has been mired in scandal. In July, Ava Tyson left the company following accusations that she shared inappropriate sexual messages with minors. In a memo obtained by the Associated Press, Donaldson revealed that he had hired a law firm to investigate Tyson. The backlash has lingered: On his X post announcing the new video, two top comments accuse him of “covering up for pedophiles” and befriending “a groomer.”

Meanwhile, the YouTuber’s ambitious “Beast Games” reality competition, which touted a $5 million grand prize, had its own complications. Contestants went online to claim injuries sustained during the games, as well as inadequate access to food, water, and medication. For a separate video, “We Schooled Hundreds of Teachers,” MrBeast listed Stand Together as a key contributor. That nonprofit happens to be founded by right-wing billionaire Charles Koch.

MrBeast’s brand of dramatic, expensive videos now has dozens of copycats, proof of his ingenuity. But, with waning growth and growing claims of wrongdoing, MrBeast’s reign of YouTube dominance could be under threat.

Democratic presidential candidate Kamala Harris defended the Biden administration’s handling of illegal immigration in a combative television interview on Wednesday, blaming Republicans for failing to pass a border security bill.

Grilled by host Bret Baier on Fox News, Harris also defended President Joe Biden’s mental fitness, her years as his vice president and her previous support of gender-affirming surgery for transgender inmates.

Harris and Baier frequently talked over each other and Harris grew visibly frustrated, but she delivered her message for the Nov. 5 election to a conservative audience that might not often hear it.

She was asked to defend the administration’s early decision to reverse some of the restrictive border policies of Republican rival Donald Trump when he was president and to respond to a mother who testified in Congress about the loss of her daughter at the hands of an immigrant in the U.S. illegally.

“I’m so sorry for her loss, but let’s talk about what is happening right now,” Harris said. She said Trump told Republicans to reject a bipartisan immigration bill early this year because “he preferred to run on a problem instead of fixing a problem.”

Trump and Republicans have claimed that immigrants are fueling violent crime in the United States, although studies show immigrants commit crimes at lower rates than others.

Asked about her recent comment that there was “not a thing” she would change about the actions of the Biden administration, Harris said, “Let me be very clear, my presidency will not be a continuation of Joe Biden’s presidency.”

She said she would bring in new ideas from Republicans and business leaders to solve housing shortages and expand small businesses.

Trump team calls interview “train wreck”

Harris was a vocal supporter of Biden when he faced mounting questions about his mental fitness after a disastrous June debate with Trump, before dropping out of the race in July. She was asked to defend those statements.

Biden has the “judgment” and “experience” to be president, she said, while questioning Trump’s fitness for office. “Joe Biden is not on the ballot, and Donald Trump is,” Harris said.

She was pressed on her position on using taxpayer funds for gender-affirming surgery for transgender inmates, including those who are undocumented. Trump has spent millions of dollars in ads on the subject in battleground states.

“I will follow the law,” Harris said, noting that the U.S. Bureau of Prisons provided gender-affirming treatments under Trump. She accused him of “throwing stones when you live in a glass house.”

The nearly 30-minute interview marked the first time Harris has appeared as a presidential candidate on the conservative media network, which often features opinion show hosts who mock her and other Democrats and tout Trump’s policies.

Within minutes of the interview’s end, the Trump campaign released a statement calling it a “train wreck.”

David Urban, a political strategist and past Trump campaign aide, said Harris gave an uneven and subpar performance, avoiding responsibility and making Trump the scapegoat. “Another losing media cycle for the Harris campaign,” Urban said.

Democrats said Harris went on unfriendly territory and made it through without any gaffes. Repeated interruptions kept Harris’s answers short, they said, preventing the meandering answers that she has been criticized for in the past.

“We feel like we definitely achieved what we set out to achieve,” said Brian Fallon, a Harris spokesperson. “She was able to reach an audience that has probably been not exposed to the arguments she’s been making on the trail, and she also got to show her toughness in standing tall against a hostile interviewer.”

Courting Republicans

The interview was part of a direct appeal by Harris on Wednesday to Republican voters. Before the Fox News interview, she highlighted Republican support for her campaign in a pivotal county in Pennsylvania, one of the handful of swing states likely to determine the election.

In Bucks County outside of Philadelphia, Harris emphasized Trump’s attempt to overturn his election loss four years ago, when he lost the White House to Biden.

She said Trump’s actions violated the U.S. Constitution and that, if given the chance, he would violate it again.

“He refused to accept the will of the people and the results of a free and fair election. He sent a mob, an armed mob, to the United States Capitol, where they violently assaulted police officers, law enforcement officials and threatened the life of his own vice president,” Harris said.

More than 100 Republicans joined Harris in Bucks County, including Adam Kinzinger, a former congressman and member of the committee that investigated the Jan. 6 attack on the Capitol by supporters of then-President Trump.

“No matter your party, no matter who you voted for last time, there is a place for you in this campaign,” Harris said. She led Trump by a marginal 46% to 43% in a recent Reuters poll.

Trump on Wednesday participated in a town hall for Latino voters hosted by the Spanish-language network Univision. Fox News aired another Trump town hall with an all-women audience.

Harris has previously sought to court voters disillusioned by Trump. former Republican congresswoman Liz Cheney, daughter of former Vice President Dick Cheney, this month urged voters to put country over party and vote for Harris, saying Trump was not fit to lead the U.S.

Biden beat Trump in Bucks County by about 17,000 votes in the 2020 election, while former Secretary of State Hillary Clinton beat Trump there in 2016 by fewer than 3,000 votes, according to county data.

This summer, Republicans overtook Democrats in the number of voter registrations in Bucks County for the first time in a generation. Republicans now have some 3,500 more voters in the county than Democrats, according to the latest data.

Harris is also weighing joining the podcaster Joe Rogan, whose show reaches millions of men across the political spectrum, and who has joked that a “puppet master” was behind Harris’s strong debate performance against Trump last month.

—Nandita Bose, Stephanie Kelly, and Jarrett Renshaw, Reuters

With the recent release of the Netflix drama Monsters: The Lyle and Erik Menendez Story and the follow-up documentary The Menendez Brothers, true-crime enthusiasts are now turning their attention to the Beverly Hills mansion where it all went down. 

Posts of the house’s exterior and interior have been going viral on TikTok, gaining millions of views from those invested in the case. One video with 1.4 million views, posted on October 6 by user @manizzle_16, sees the user drive by 722 North Elm Drive in Los Angeles. Another video with more than 818,000 views, from a Menendez brothers fan account, shows a slideshow of photos of the house dating back to the 1980s, including the room where the murders took place.

On September 25, user @momonii_ posted a video they captioned, “Come with me to look inside the house where The Menendez brothers killed their parents.” Viewed over 1.1 million times, we see the TikToker walk right up to the front door and hold her phone up to the glass so that curious fans can catch a glimpse of the inside. 

However, in another video, user @michaelpavano explains that people are getting the actual house confused with its neighbor. The Menendez mansion, recently purchased in March, is currently undergoing renovations and, as a result, is surrounded by a high, screened construction fence. In the video, with 398.1K views, he holds his phone over the fence to give viewers a glimpse of the real Menendez house. “I’d be lying if I said I wasn’t tempted to hop the fence and film something for you guys,” he says. “But that would be incredibly disrespectful and highly illegal. So obviously, I’m not gonna do that.” The famous house has changed hands many times in the 35 years since the murders, most recently selling for $17 million.

TikTok clips are often soundtracked to either “Girl I’m Gonna Miss You” or “Blame It on the Rain” by Milli Vanilli, both of which appear in Monsters. Following the nine-episode series’s release, TMZ reported that the property has become a magnet for true-crime enthusiasts, with locals noticing an uptick in visitors hoping to see it firsthand.

Lyle and Erik Menendez were found guilty of murdering their parents in 1989 and have spent nearly 35 years behind bars. On October 3, Los Angeles District Attorney George Gascón revealed that his office would reexamine new evidence provided by the brothers’ lawyer, aiming to support their claim that the killings resulted from years of sexual abuse by their father and fear for their own safety.

Meta Platforms laid off employees from multiple teams, including Instagram, WhatsApp, and Reality Labs, in a fresh wave of job cuts at the Facebook parent company on Wednesday. 

“A few teams at Meta are making changes to ensure resources are aligned with their long-term strategic goals and location strategy,” said a Meta spokesperson in an emailed statement to Fast Company. “This includes moving some teams to different locations, and moving some employees to different roles. In situations like this when a role is eliminated, we work hard to find other opportunities for impacted employees.”

Among the affected employees is Jane Manchun Wong, who took to Meta-owned Threads to announce the news. “I’m still trying to process this but I’m informed that my role at Meta has been impacted,” Wong wrote in a public Threads post on Wednesday.

Some employees were given the option of taking a different position under a new contract, or taking a severance package, with several choosing the latter, an anonymous former Meta employee told TechCrunch.

Meta declined to provide details on the number of employees affected. 

Zuck’s “Year of Efficiency” continues 

Meta has experienced multiple rounds of layoffs following aggressive hiring during the pandemic. It first announced that it was reducing its workforce back in November 2022, a notable change of direction at the social-networking giant after years of head-spinning growth. It said at the time that it would lay off about 11,000 people.

Layoffs continued the following year, with CEO Mark Zuckerberg proclaiming 2023 the “Year of Efficiency” (a strategy also adopted by other Big Tech companies after pandemic-era hiring) and announcing another 10,000 layoffs.

And earlier this year, Meta’s Reality Labs division, which produces virtual reality and augmented reality hardware and software, also saw a small series of job cuts.

Last week, Meta also reportedly fired about two dozen staff in Los Angeles for routinely misusing $25 meal credits to buy non-food items over an extended period of time, according to the Financial Times.

For decades, chief executive pay has climbed precipitously, buttressed by generous stock awards and six-figure bonuses. Since 1978, CEO compensation has spiked by 1,085%, a far cry from the 24% bump in pay that the average worker has seen.

In 2023, however, overall CEO compensation—which includes salary, bonuses, stock awards, and stock options—actually dropped by nearly 20%, according to a recent analysis by the Economic Policy Institute (EPI). That means CEOs were only paid 290 times as much as the average worker, a sharp decline from previous years. (Just a year before, in 2022, CEOs received 360 times the pay of the average worker.)

As the Washington Post reports, however, it’s not clear why this is the case, since chief executive pay usually mirrors stock market performance. EPI researchers have said it’s unusual to see a dip in CEO compensation during a year when the stock market was strong—which means this could either be an anomaly, or perhaps indicate a broader shift in how CEOs are paid.

The EPI analysis also notes that it could just be that CEOs are receiving more stock awards, rather than options, which could be “a promising move to align CEO pay to longer-term incentives.” (Another report by Equilar earlier this year found that median total compensation for CEOs increased by 12.6% to $16.3 million, though its methodology and criteria differ from the EPI analysis.)

It seems unlikely, however, that this signals a more sweeping change, given how steadily CEO compensation has increased over the years, even amid growing outrage over bloated pay packages. As Fast Company noted in a special report on CEO pay last year, unions and politicians have drawn attention to the yawning pay disparities between employees and CEOs, particularly as more workers have gone on strike.

Fast Company‘s analysis of CEO pay—conducted in partnership with MyLogIQ—found that some of the “least fairly paid” chief executives had pay packages worth many thousands of times the median pay for their workers. Even shareholders have been realizing that sky-high CEO compensation can be a liability, as it invites greater scrutiny and regulation in the interest of more transparency. But the question is whether mounting public pressure will actually move the needle on CEO compensation—or if it’s already too far gone.

Abortion will be on the ballot in 10 U.S. states for the Nov. 5 election, including battleground states that will play critical roles in the presidential race and the fight for control of Congress.

Democrats, led by Vice President Kamala Harris, the party’s presidential candidate, have sought to build support for the measures, which they hope will galvanize left-leaning and independent voters. Harris faces Republican former President Donald Trump in the election.

The states where voters will decide whether to guarantee abortion rights include Arizona and Nevada, which are not only among the swing states likely to decide the presidential contest but also feature high-profile Senate races that could determine which party controls that chamber.

The issue of abortion has bedeviled Republicans since the U.S. Supreme Court’s 2022 decision to eliminate a nationwide right. Anger over the ruling was widely credited with limiting Republican gains in the 2022 midterm elections, as well as boosting Democrats’ performance in state races in Kentucky and Virginia last year.

Voters have chosen to protect or expand abortion access in all seven statewide ballot measures put to a vote since the court’s decision, including in conservative strongholds such as Ohio, Kentucky and Kansas.

Here are the states where abortion will be on the ballot on Nov. 5.

Arizona

Arizona election officials approved a ballot measure that would amend the state constitution to guarantee abortion rights up to fetal viability, generally considered to be around 23 or 24 weeks.

Abortions in Arizona are currently subject to a 15-week ban that Republican lawmakers passed in 2022 after the U.S. Supreme Court decision.

President Joe Biden narrowly won the state over Trump in 2020, the first Democrat to carry it in 24 years, and polls show it is once again a close contest between Harris and Trump.

The state’s race for an open Senate seat between Democrat Ruben Gallego and Republican Kari Lake is one of a handful expected to determine which party controls the Senate.

Florida

The state Supreme Court on April 1 approved a ballot measure, backed by reproductive rights groups, asking voters whether to amend the state constitution to protect abortion access.

As of May 1, abortions in Florida are banned after six weeks with only rare exceptions under a law backed by Republican Governor Ron DeSantis.

Unlike in most states, constitutional amendments in Florida must pass with at least 60% of the vote, a higher threshold of support than any statewide abortion measure has yet received. A New York Times/Siena College poll released in early October found the measure falling short, with 46% of voters in support, 38% opposed and 16% undecided.

Once a perennial battleground state, Florida has leaned Republican in recent elections, voting twice for Trump and reelecting DeSantis in a landslide in 2022.

Trump, a Florida resident, has said he will vote against the ballot measure, after initially appearing to suggest he would vote in favor.

Nevada

State officials approved a referendum for November that would amend the state constitution to protect abortion rights, after a coalition of reproductive rights groups submitted enough valid signatures.

State law already offers similar protections, but adding them to Nevada’s constitution would make it harder to roll those rights back. Voters would need to approve the measure twice – this year and again in 2026 – to amend the constitution.

Nevada is a battleground state in the presidential contest, while the Senate race between incumbent Democrat Jacky Rosen and Republican challenger Sam Brown is one of the most closely watched of 2024.

Other states

Voters in Missouri and South Dakota – both deeply conservative states where virtually all abortions have been banned – will also decide whether to add abortion rights to their state constitutions.

Nebraska voters will face two opposing ballot measures. One would add the right to access abortions to the state constitution, and the other would enshrine the state’s current 12-week ban – with exceptions for rape, incest and the life of the mother – in the constitution.

In the unlikely event that both measures pass, the one that gets the most affirmative votes would take precedence, according to state officials.

Montana and Colorado, two states where abortion has remained legal, also have ballot measures that would add abortion rights to their state constitutions. Under Colorado law, ballot measures need 55% support to pass.

Supporters say adding protections to those states’ constitutions would ensure lawmakers or courts could not limit rights in the future. Montana’s Republican-controlled legislature has sought to overturn the state Supreme Court’s 1999 ruling upholding the right to abortion.

While not a presidential battleground, Montana has a highly competitive U.S. Senate race this year between Democratic incumbent Jon Tester and Republican challenger Tim Sheehy.

Lawmakers in two other states have approved abortion-related amendments for November’s ballot: New York and Maryland. In both states, abortion is already legal; the referendums would amend their state constitutions to add more protections.

—Joseph Ax and Gabriella Borter, Reuters

Additional reporting by Liya Cui.

On Tuesday, Congresswoman Rashida Tlaib took to X with a letter directed at Rodney McMullen, the CEO of Kroger. In it, she expressed her concern about Kroger’s electronic shelving labels (ESLs), as well as the grocery chain’s alleged plans to use facial recognition technology at its digital displays—both tools, she wrote, have the potential to be “abused in the pursuit of profit.”

Tlaib isn’t the first U.S. official to voice her trepidation at Kroger’s ESL tech, either. Back in August, Senators Elizabeth Warren and Robert Casey wrote to McMullen, stating that they were “concerned about whether Kroger and Microsoft are adequately protecting consumers’ data” and laying out an extensive list of questions about the company’s plans for its ESLs (which, as of this writing, Kroger has not publicly responded to). 

Both letters have sparked concern—and a healthy dose of shock—from X users, many of whom shared that they had no idea such technology existed in grocery stores. But despite the fact that public concern is surfacing now, Kroger has actually been working on its ESLs for at least five years, and tells Fast Company it has no plans to use facial recognition in its stores.

Here’s everything we know about Kroger’s ESLs, and what they might mean for consumers.

What is an ESL?

An electronic shelving label is essentially the digital version of the price tag found throughout grocery stores. For companies, these labels are beneficial because they cut down on the manual labor of updating the tags, can help stores quickly update prices on items and keep track of their inventory, and can even help workers and shoppers identify items through a flashing light system. Some stores are also using the tags to display barcodes with useful information, such as nutrition facts. 

In an interview with Grocery Dive, Anne Mezzenga, co-CEO of retail blog Omni Talk, explained, “They tackle the big issue of changing prices and the loss that grocery retailers can experience when they don’t have accurate pricing on shelf.” She added that ESLs can help cut down on waste by eliminating the need to print out tags and shelf talkers (engaging little signs posted on a shelf edge further promoting a product).

Which grocery stores are using ESLs, and when did they start?

Kroger began using ESLs back in 2018 and currently uses ESLs in at least 500 locations. This June, Walmart announced that it would be rolling out ESLs to 2,300 stores over the next two years after a successful pilot. 

Other stores, including Schnucks, Hy-Vee, Ahold Delhaize USA, and Dom’s Kitchen & Market, have also embraced the technology in recent years. 

Why are people concerned about ESLs?

While industry leaders continue to espouse the usefulness of ESLs, government officials and media are raising several red flags around how the tech might become harmful. 

The first issue is the fear that stores might start using ESLs for dynamic or surge pricing. This describes the idea that, in order to turn a higher profit, grocery stores could start adjusting prices depending on time of day, demand, or other factors—similar to something like Uber’s pricing model.

In their letter to Kroger, Warren and Casey wrote, “For example, stores may use this technology to raise the price of turkeys in the days leading up to Thanksgiving, or the price of ice cream on a hot day, causing customers to face unexpected price hikes when they reach the grocery aisle and find themselves unable to afford the groceries that they had originally budgeted for.”

For their part, both Kroger and Walmart have assured customers that they will not use ESLs for dynamic pricing, and will instead employ the tech to offer shoppers better discounts on items like final-clearance goods, for example.

“Kroger’s business model is built on a foundation of lowering prices to attract more customers,” a Kroger spokesperson told Fast Company. “Everything we do is designed to support this strategy, and customers are shopping more with Kroger now than ever because we are fighting inflation and providing great value. To be clear, Kroger does not and has never engaged in ‘surge pricing.’ Any test of electronic shelf tags is designed to lower prices for more customers where it matters most. To suggest otherwise is not true.”

What about facial recognition technology

Another concern Tlaib raised in her letter was that grocery stores might start using facial recognition technology as part of their in-store advertising and pricing efforts. In her letter to Kroger, Tlaib wrote, “The use of facial recognition tools has the potential to invade a customer’s privacy and employ biased price discrimination.”

In 2019, Fast Company explored this issue in an article on an ESL tech. We reported that Kroger was piloting a new technology in one of its stores, called the Enhanced Display for Grocery Environment (EDGE). That year, Kroger told Fast Company that it was teaming up with Microsoft to implement AI-powered facial recognition technology in some of its digital displays.

The companies announced in a 2019 press release, “the EDGE Shelf will enable Kroger to generate new revenue by selling digital advertising space to consumer packaged goods (CPGs) brands. Using video analytics, personalized offers and advertisements can be presented based on customer demographics.” 

The proposition, as described to Fast Company, was to use gender and age information to display personalized ads or coupons. Tlaib, Warren, and Casey cited Fast Company’s 2019 article as their main evidence for the idea that Kroger plans to use facial recognition and “use customer data to build personalized profiles of each customer.” 

However, according to a Kroger spokesperson, the one-store EDGE pilot described in that 2019 coverage was quickly sunset, and Kroger no longer uses EDGE technology in any of its stores. The spokesperson added that Kroger is not, and has never, used facial recognition in its stores—nor are they aware of any plans to do so in the future.

Federal disaster personnel have resumed door-to-door visits as part of their hurricane-recovery work in North Carolina, an effort temporarily suspended amid threats that prompted officials to condemn the spread of disinformation.

Over the weekend, reports emerged that workers with the Federal Emergency Management Agency could be targeted by militia as the government responds to Hurricane Helene. A sheriff’s office said Monday that one man was arrested during an investigation, but that the suspect acted alone.

FEMA made operational changes to keep personnel safe “out of an abundance of caution,” agency administrator Deanne Criswell said at a briefing Tuesday. FEMA workers were back in the field Monday, accompanied by Criswell, and she said disaster-assistance teams helping survivors apply for FEMA aid as well as state and local assistance will continue to go door-to-door. She emphasized that the agency isn’t going anywhere.

“The federal family has been here working side by side with the state since day one. These are people who put their lives on hold to help those who have lost everything,” Criswell said. “So let me be clear: I take these threats seriously.”

North Carolina Governor Roy Cooper said he directed the state’s Department of Public Safety to coordinate law enforcement assistance for FEMA and other responders. He stressed the damage that internet rumors and falsehoods were causing and said officials may never know how many people won’t apply for assistance because of bad information.

“There’s still a persistent and dangerous flow of misinformation about recovery efforts in western North Carolina that can lead to threats and intimidation, breeds confusion, and demoralizes storm survivors and response workers alike,” Cooper said at the briefing. “If you’re participating in spreading this stuff, stop it. Whatever your aim is, the people you are really hurting are those in western North Carolina who need help.”

The Rutherford County Sheriff’s Office said it received a call Saturday about a man with an assault rifle who made a comment “about possibly harming” FEMA employees working in the hard-hit areas of Lake Lure and Chimney Rock in the North Carolina mountains. A man was charged with “going armed to the terror of the public,” a misdemeanor, and was released after posting bond.

The sheriff’s office said it received initial reports that a “truckload of militia” was involved in the threat, but further investigation determined the man acted alone.

FEMA has faced rampant disinformation about its response to Helene, which hit Florida on September 26 before heading north and leaving a trail of destruction across six states.

Asked what might be fueling disinformation, Cooper said social media has become more extreme, but he also pointed to politics.

“This is happening in the middle of an election where candidates are using people’s misery to sow chaos for their own political objectives—and it’s wrong,” he said.

Former President Donald Trump and his allies have seized on the storm’s aftermath to spread false information about the Biden administration’s response in the final weeks before the election. Their debunked claims include false statements that victims can receive only $750 in aid, that emergency response funds were diverted to immigrants, that people accepting federal relief money could see their land seized, and that FEMA is halting trucks full of supplies.

Helene decimated remote towns throughout Appalachia, left millions without power, knocked out cellular service, and killed at least 246 people. It was the deadliest hurricane to hit the U.S. mainland since Katrina in 2005.

Terrie Daughtry, a volunteer handling therapy dogs Tuesday at a FEMA Disaster Recovery Center in Asheville, said threats and misinformation—including the militia rumors—made her feel unsafe for the first time in several trips to volunteer at disaster sites.

“I’m not coming to risk my life with it all, to be shot or hurt or trampled because of lunacy,” said Daughtry, who volunteers with Therapy Dogs International. She said she previously traveled to help in the aftermath of the Pulse Nightclub shooting in Orlando, Florida, floods in Virginia, and tornadoes in Texas, Oklahoma, and Alabama.

She and another volunteer have been using their therapy dogs to calm people waiting in line to make FEMA claims. They hand out candy, let people pet the dogs, and talk to people about their experiences.

Despite the extra stress from the “absolutely ridiculous” threats and misinformation, Daughtry said she’s seen some amazing moments of human spirit. At one point on Monday, someone in line started playing a guitar and singing about having no water, she said. Eventually, the whole line sang along.

“These are special people. They’re singing in horrible adversity,” she said. “It made me tear up being there and it’s making me tear up now.”

—By Makiya Seminera and Sarah Brumfield, Associated Press

Deepfakes have already been a problem in the 2024 presidential election—and could potentially become a bigger one as November 5 draws near. But a new Chrome browser extension claims it can help people determine which audio clips they hear are legitimately from the candidates (and others) and which are fake.

The Hiya Deepfake Voice Detector uses artificial intelligence to determine if the voice on screen is legitimate or faked. The company claims the tool has a 99% accuracy rate and says it can verify (or debunk) audio in just a few seconds.  

“Our models are trained to detect subtle audio artifacts unique to AI-generated voices—imperceptible to the human ear but identifiable by machine learning algorithms—and just a second of audio is enough to detect their presence,” Patchen Noelke, vice president of marketing for Hiya, tells Fast Company.

I put the free tool to the test and found that while it certainly does catch some deepfakes, its accuracy varied.

Testing the browser extension

Founded in 2015 as part of online directory Whitepages, Hiya is in the business of providing services that screen calls for spam and fraud. Today, the company has more than 450 million active users, according to its website.

After installing the Hiya’s extension from the Chrome Web Store and registering, users open the tool in their browser and click, “Start analyzing,” when they’re watching a video or listening to an audio clip whose origin is in doubt. Users are limited to 20 queries per day. 

While it’s certainly helpful to have any sort of tool that identifies even a percentage of the deepfake video and audio clips (others, though not browser extensions, include Pindrop Security, AI or Not, and AI Voice Detector), my own tests with the Hiya Google Chrome app found some notable holes.

Using real and deepfake soundbites that appeared in a Washington Post story, I had the tool examine both a real and a fake clip of Kamala Harris. To its credit, it did call out the deepfake, but the tool said it was uncertain about the authentic clip of Harris.

An AI-generated audio file of Trump was judged as likely authentic, with an authenticity score of 88 out of 100. And a confirmed deepfake of Ukrainian president Volodymyr Zelenskyy instructing his troops to lay down their weapons got an even higher authenticity score of 90.

Deepfakes of Mark Zuckerberg and an AI-created video that featured a voice actor imitating Morgan Freeman were also wrongly deemed authentic, with the fake Zuckerberg earning a score of 79 and the fake Freeman getting a 72.

A deepfake of Kim Joo-Ha, a news anchor on Korean television channel MBN (which was presented by the network last year), was spotted immediately as a fake, however, earning a score of just 1.

Noelke acknowledges that the deepfake detector has some limitations. Freeman and Zelenskyy fakes were made with voice actors rather than AI, which is why they weren’t detected, he says. The score on the Trump deepfake changed depending on how much audio was sampled, Noelke notes. (“When scoring the entire file instead of analyzing it in 4-second chunks, the overall score is 63%, placing it in the inconclusive category,” he says.) And the Harris deepfake had loud background music that interfered with scoring, which is why the tool was uncertain.

“The way we designed the Chrome extension is it gives you a confidence score and you do have to judge for yourself,” he says. “There are some things that make it harder [like] the quality of the sound and background music and other things can affect it. It can definitely fail. It’s definitely not 100%—and that’s why we deliver the results the way we [do].”

The app also only analyzes the audio portion of a video, so if a politician, celebrity, or anyone else’s image is digitally inserted into a video where they do not actually appear, it cannot flag that as inauthentic. We tested this on a Back to the Future clip that swaps Tom Holland and Robert Downey Jr. for Michael J. Fox and Christopher Lloyd.

Deepfakes and the election

AI has already thrown a wrench in the 2024 presidential election. In January, a deepfake robocall that reproduced the voice of President Joe Biden urged Democrats not to vote in the New Hampshire primary. Since then, there have been false images that show everything from Taylor Swift endorsing Trump to Kamala Harris in a communist uniform to Donald Trump dancing with an underage girl. And this summer, Elon Musk shared a deepfake that mimicked the voice of Harris claiming to be a “diversity hire” and saying she didn’t know “the first thing about running a country.”

Musk, when criticized, later said the video was intended as satire.

Several states have enacted laws regulating election deepfakes, including California, Texas, and Florida. But only three of the eight swing states that are expected to help decide the 2024 presidential election have laws on the books—Michigan, Wisconsin, and Arizona. (Legislation is pending in North Carolina and Pennsylvania but is unlikely to pass before the election.)

The Federal Communications Commission (FCC) quickly made AI-generated voices in robocalls illegal, but federal officials have been slower to act on other mediums. And when it comes to online content, which is where these most frequently spread, that continues to be the wild west. The FCC says it does not regulate online content. 

Update, October 17, 2024: This story has been updated with a response from Noelke about the specific audio clips tested.

Polls opened for early voting on Thursday in North Carolina—a critical battleground in the 2024 U.S. presidential election—just weeks after Hurricane Helene left some residents still without power, running water, or phone service.

North Carolina could be crucial in determining a winner in the race between Vice President Kamala Harris, a Democrat, and former Republican President Donald Trump because its voting preferences could swing either way. Polls show a very tight race with a handful of swing states likely to determine the winner.

Hurricane Helene, which killed more than 200 people across six states and caused billions of dollars in damage, has sparked questions on how it might impact voter turnout.

Harvey Neal, a 71-year-old retiree, voted at the lone early voting site in Hendersonville, North Carolina, as a steady stream of voters filed in and out of the center on Thursday morning.

Neal cast his vote for Harris. “She’s young, she’s a Democrat,” he said. “And, you know, what’s the alternative?”

Neal considers himself relatively fortunate following the hurricane as nothing hit his house, though he had been without power for four or five days.

Trump and Harris are in a dead heat in North Carolina, with Trump showing 48% support in state polls, versus 47.5% for Harris, according to poll aggregator FiveThirtyEight.

Some of the state’s hardest-hit counties lean both Republican and Democratic. For instance, Buncombe County—home to the city of Asheville—supported President Joe Biden in the 2020 election versus Trump, while neighboring Yancey County favored Trump.

North Carolina’s state elections board has spent weeks assessing storm damage to polling locations. On Thursday, 76 early voting sites were set to open to voters in the 25 western North Carolina counties listed in the federal disaster declaration. That compares to 80 planned before the storm.

Early voting was the most popular way for North Carolinians to cast a ballot in both the 2020 and 2016 elections, according to the state. Early voting started on Tuesday in Georgia, another closely contested state that was hit by storm damage.

Tina Veitch, a 49-year-old graphic designer, said the road to her family’s house in Burnsville was destroyed by the storm. But one of the reasons she has stayed in the area, instead of seeking refuge with family in Florida, is so she can vote.

She said on Wednesday that the process to vote by absentee ballot was too arduous, adding: “The amount of hoops you have to jump through to just do it was not worth it. . . . So we were like, ‘We’re just going to stay ’til tomorrow at least and vote.’”

—By Stephanie Kelly, Reuters

Additional reporting by Trevor Hunnicutt.

Meta has sacked a number of staff for abusing the company’s $25 meal scheme to order everything but lunch—including acne pads, toothbrushes, and wine glasses. 

The sacking of nearly 30 staff in the company’s Los Angeles office, first reported by the Financial Times, included high-paid engineers earning six-figure salaries. 

One user of the anonymous chat app Blind wrote that they used the credits for “non-food items, shared credits with people, or went above budget.” Other examples of the nonfood items bought included toothpaste and assorted household items. “They were given a warning to stop, which most of them did, but were still fired three months later even after stopping,” the user said.

Another user reportedly wrote: “On days where I would not be eating at the office, like if my husband was cooking or if I was grabbing dinner with friends, I figured I ought not to waste the dinner credit.”  According to the Financial Times, the worker confessed to the misuse of company credits after being investigated by human resources, and was subsequently fired. 

“It was almost surreal that this was happening,” the person wrote.

Mark Zuckerberg’s Meta, which is currently worth $1.5 trillion, offers staff free breakfast, lunch, and dinner at its larger offices as a company perk. The social media giant’s Penn Station office in New York City, for example, features a cafeteria akin to an upscale food court. 

However, those in smaller offices without staff canteens are instead supplied with vouchers for delivery apps such as Grubhub to use while working at the office. However, some have been caught taking full advantage of the program, using the credits to have dinner delivered to their homes or pooling their funds with colleagues. Those who were let go were repeat offenders of misusing the vouchers, while others were reprimanded but not fired. 

Meta disclosed in a regulatory filing earlier this year that the median total annual compensation for its employees, excluding CEO Mark Zuckerberg, is $379,050. News of the firings follows Meta’s announcement on Thursday that it is cutting jobs across the company as part of several separate restructurings.

“Today, a few teams at Meta are making changes to ensure resources are aligned with their long-term strategic goals and location strategy,” Meta spokesperson Tracy Clayton said in a statement. “This includes moving some teams to different locations, and moving some employees to different roles. In situations like this when a role is eliminated, we work hard to find other opportunities for impacted employees.”

With just 19 days until the 2024 presidential election, early voting returns have become a major focus for both candidates as Americans cast pre-Election Day ballots in record numbers.

What is early voting?

Early voting varies by state. Some states have already started early voting, either in person, by mail, or both. (For example, in my home state of Massachusetts, voting by mail has already started, and in-person early voting starts this Saturday.)

How many people have voted so far?

The Election Lab at the University of Florida tracks early voting. Here’s the latest count from this tracker as of Thursday afternoon:

• 8,336,501 : total early votes
• 1,814,362 : in-person early votes
• 6,489,081 : mail ballots returned  
• 55,491,644 : mail ballots requested

What about early voting in the battleground states?

Here’s a look at early voting so far in the seven swing states—Georgia, Arizona, Pennsylvania, Wisconsin, Michigan, North Carolina, and Nevada—which could determine the outcome of the 2024 election.

Residents in the battleground State of Georgia have already cast more than 620,000 votes, according to the University of Florida Election Lab tracker.

Meanwhile, in Arizona, where Democratic nominee Vice President Kamala Harris and Republican candidate former President Donald Trump have been campaigning hard, more than 77,000 residents have already voted. This is worth watching in the run-up to November 5 because nearly 90% of Arizona voters cast their ballots early in the 2020 election.

And in the hotly competitive state of Pennsylvania, more than 629,000 residents have cast their ballots so far.

Among the states that report party affiliation (not all do), here’s how the votes break down by party:

• 51.4% Democrats
• 30.5% Republicans
• 18.1% Minor/No Party Affiliation

You can check out the full numbers and interactive map on the University of Florida’s Election Lab website, run by political science professor Michael McDonald, who has been tracking early voting since 2008.

Republican presidential candidate Donald Trump says that if he does not win the Nov. 5 election, he will cry fraud and not accept the results – just as he did four years ago when he lost to Democratic President Joe Biden.

“If I lose – I’ll tell you what, it’s possible. Because they cheat. That’s the only way we’re gonna lose, because they cheat,” Trump said at a Michigan rally in September.

A refusal by Trump to accept a victory by Democratic rival Kamala Harris could throw the United States into political instability at a time when the country is already deeply divided.

After Trump lost the 2020 election, he and his allies attempted to overturn the result through dozens of lawsuits that ultimately failed to alter or delay the vote count.

He also pressured officials in Georgia to find more votes for him; and his supporters stormed the U.S. Capitol on Jan. 6, 2021 in a failed effort to stop his vice president, Mike Pence, from certifying Biden’s victory.

One key difference this time is that Trump does not have the presidential levers of power that he did in 2020. And new state and federal laws have been put in place to make it more difficult to influence election results.

Still, Trump and his allies have been laying the groundwork for months to cry foul if he loses on Nov. 5. He could contest a win by Harris in the courts or raise doubts about the validity of her victory among supporters that could have unforeseen consequences.

Smoke and mirrors

Republicans and Democrats expect that vote counting could drag on for several days after Nov. 5 as mail-in ballots are tabulated and other votes are tallied and verified.

If it appears Trump is losing, the delay will give him an opportunity to claim fraud and attempt to undermine confidence in election officials, while also possibly encouraging his supporters to protest. He has already threatened to jail election workers and other public officials for “unscrupulous behavior,” although he would need to win the election first.

Trump can take his case directly to the American public without waiting for proof, using social media, press conferences and interviews.

“President Trump has been very clear that we must have a free and fair election,” said Karoline Leavitt, a spokesperson for the Trump campaign.

The Harris campaign did not immediately respond to a request for comment.

In the battlegrounds

Republicans have already preemptively filed more than 100 lawsuits in the battleground states that will decide the election to seed the ground for post-election challenges, including claiming, without evidence, that non-citizens will be voting in large numbers.

Both parties plan to dispatch thousands of trained volunteers called poll watchers to monitor voting and vote counting with a mandate to report any irregularities.

Some voting rights activists are concerned that Republican poll watchers could be disruptive, but the Republican Party says the volunteers have been trained to stay within the law.

As they did in 2020, Trump’s allies in key states – local election officials, state lawmakers and perhaps judges – could seek to delay certification, the confirmation of a state’s official tally, through claims of fraud.

Those efforts did not succeed last time, and election law experts say the laws in those states are clear that local officials lack the power to throw out ballots or derail the process.

Five of the seven battleground states have Democratic governors but Democratic activists worry about Georgia, whose state election board recently gave unprecedented authority to local officials to conduct inquiries, a move they say could give an opening to bad-faith actors who attempt to contest or delay the vote count.

A Georgia judge, however, ruled this week that local officials must certify the results and do not have the discretion to do otherwise.

All states must submit their certified totals before the Electoral College meets in December and electors cast their votes. That vote is then delivered to Congress for final certification in January.

Trump-inspired court challenges and certification delays could cause a state to miss the filing deadline. That could provide grist for Republican objections in Congress.

Some election law experts caution that it is difficult to predict how novel legal disputes over certification might be resolved, especially if they are handled by judges sympathetic to Trump’s claims.

Congress has final say

After the 2020 election, Congress passed a reform law that makes it more difficult for candidate to mount the kind of challenge Trump attempted.

It makes clear that the vice president, who in this case would be Harris, has no authority to delay national certification or throw out a state’s results, as Trump urged Pence to do in 2020.

The measure also requires that an objection to a state’s electoral count cannot be brought unless one-fifth of the members of each house of Congress agrees. After that, it takes a majority vote in each house for an objection to be found valid.

In the unlikely result that enough electoral votes are tossed so that neither candidate reaches the necessary majority, the newly elected U.S. House of Representatives would choose the next president.

Civil unrest

Any effort by Trump to suggest the election was rigged could potentially lead to civil unrest, as it did on Jan. 6, 2021.

Experts who monitor militant right-wing groups, such as Peter Montgomery of the People For the American Way, a liberal think tank, say they are less concerned about a violent response from these groups than they are about threats against election workers counting votes. There also could be violent demonstrations in the capitals of battleground states, Montgomery said.

Hundreds of people who were involved in the Jan. 6 attack on the Capitol have been convicted and jailed for their actions, a powerful deterrent to others who may be considering taking similar actions.

—James Oliphant, Reuters

Additional reporting by David Morgan and Jack Queen.

A good year for Netflix is only getting better, as the streaming giant saw its ranks of paid subscribers surge 14.4% to 282.7 million members during the third quarter of 2024.

The company also reported revenue and earnings that exceeded Wall Street forecasts, with revenue up 15% in the three months ended September 30 to top $9.8 billion, and earnings per share at $5.40.

Shares of Netflix rose more than 3% in after-hours trading; the stock has surged nearly 48% so far this year. That’s more than double the year-to-date gains for the Nasdaq 100, a benchmark of largely tech companies.

In the trenches of the streaming wars

Netflix has continued to maintain its dominance in an increasingly crowded market for streaming, and is seeing big gains in its lower-priced, advertising-supported plans. The Las Gatos, California-based company reported that membership in its ads plan grew 35% from the prior quarter and accounted for more than half of sign-ups in the countries where it’s available. Netflix plans to roll out the ads plan in Canada next month, followed by a broader global rollout in 2025.

The company also said it expects paid net additions to be higher in the fourth quarter thanks to “normal seasonality and a strong content slate,” which includes the second season of Squid Game, the Jake Paul vs. Mike Tyson fight, and two NFL games scheduled to air on Christmas Day.

The United States and Canada remain Netflix’s largest market, with about 44% of revenue generated from these countries in the third quarter—a share that’s remained largely consistent over the past two years.

Trump churn proves inconsequential

Despite a widely circulated report about Donald Trump supporters ditching the streaming service over the summer, politically motivated cancellations proved inconsequential.

Netflix saw the rate of U.S. cancellations nearly triple in the days immediately after Reed Hastings, its cofounder and chairman, endorsed Vice President Kamala Harris for president on July 22 and donated millions to her campaign, according to data from Antenna and reported by Bloomberg News. July 26 marked the single worst day of cancellations for the year.

However, Netflix churn is far lower than that of its competitors, which is part of the reason the temporary spike was even noticeable. Overall, the company’s subscriber numbers continue to show an upward trend.

Circulation has been swirling lately, suggesting Netflix could be poised to increase prices on its plans soon—potentially as soon as this week—which could cause further churn. The last time the company did a price adjustment was in 2022.

Still, it seems Netflix subscribers are hooked, at least for the time being. The company reported that paid members view about two hours’ worth of content per day and that viewership increased in the first nine months of the year.

It cited recent hits including the latest season of Bridgerton, Baby Reindeer, Under Paris, and The Roast of Tom Brady, which it said attracted its largest live audience yet.

The company has branched out in recent years to shake up its programming by backpedaling on some prior promises that it would “never” venture into things like original programming, advertising-supported streaming, or live sports.

At the Fast Company Innovation Festival last month, Netflix co-CEO Ted Sarandos said the company would remain nimble and open to change. “I think we should always reserve the right to get smart,” Sarandos said at the September conference.

Facebook owner Meta said on Thursday it had teamed up with Blumhouse Productions, the Hollywood company behind popular horror films such as The Purge and Get Out, to put its new generative AI video model Movie Gen to the test.

The announcement comes after Meta unveiled Movie Gen, which it said can create realistic-seeming video and audio clips in response to user prompts, earlier this month. Meta claimed the tool could rival offerings from leading media generation startups such as OpenAI and ElevenLabs.

In a blog post, the social media company said Blumhouse had selected filmmakers Aneesh Chaganty, the Spurlock Sisters, and Casey Affleck to try out Movie Gen and use clips generated by the tool in their short films.

Chaganty’s film would appear on Meta’s Movie Gen website, while films from Affleck and the Spurlock Sisters were forthcoming, Meta said.

Blumhouse CEO Jason Blum said in a statement that artists remain the lifeblood of the industry and innovative technology can aid in their storytelling.

“We welcomed the chance for some of them to test this cutting-edge technology and give their notes on its pros and cons while it’s still in development,” Blum said. “These are going to be powerful tools for directors, and it’s important to engage the creative industry in their development to make sure they’re best suited for the job.”

With the partnership, Meta is signaling how it aims to collaborate with creative industries, whose members have largely recoiled in response to the arrival of generative AI technologies over concerns around copyright and consent.

Several groups of copyright owners have sued major tech companies, including Meta, over the unauthorized use of their works to train generative AI systems. Meta has argued its AI training is protected by the copyright doctrine of fair use.

However, Meta and other tech companies have shown they are willing to pay for certain types of content for AI. Aside from the Movie Gen partnership, Meta said last month it had struck deals with actors including Judi Dench, Kristen Bell, and John Cena to give voice to its Meta AI chatbot.

Similarly, Microsoft-backed OpenAI has been meeting with Hollywood executives and agents this year to discuss possible partnerships involving its video generation tool Sora, which it first showed off in February.

No deals have been reported to have come out of those talks yet, though Lions Gate Entertainment said in September that it had struck a deal with another AI startup, Runway.

—By Katie Paul, Reuters

Additional reporting by Dawn Chmielewski.

The USDA’s Food Safety and Inspection Service has released a preliminary list of schools that received ready-to-eat meat and poultry products that were recalled over fears of Listeria contamination.

The products were produced by a BrucePac processing facility in Durant, Oklahoma, and sold nationwide under familiar brand names such as Trader Joe’s, Jenny Craig, Atkins, Kroger, 7-Eleven, and many more.

The list contains more than 200 schools across 16 states and the District of Columbia. The states are:

• Delaware
• Florida
• Georgia
• Illinois
• Indiana
• Iowa
• Kentucky
• Michigan
• Missouri
• New Jersey
• New York
• Ohio
• Pennsylvania
• Tennessee
• Wisconsin
• West Virginia

The list (available here as a PDF file) does not say which schools received which specific products. The recall was first announced last week and has since been expanded to include almost 12 million pounds of product.

Earlier this week, the USDA said it had discovered that schools were among the institutions that received the products.

A spokesperson for the USDA told Fast Company that meat and poultry facilities are inspected during every shift. In this case, the Listeria contamination was discovered during routine testing of the finished product. The products sent to schools were not part of the USDA’s National School Lunch and Breakfast Programs, as schools often source food from other vendors.

BrucePac has released its own statement noting that no illnesses associated with this recall have been reported.

The FSIS says it will continue to update both the school and retail product lists as more information becomes available.

Listeria infection, or listeriosis, can cause symptoms like fever, muscle aches, headache, stiff neck, confusion, loss of balance, and convulsions, along with diarrhea and other gastrointestinal issues. It can lead to miscarriages or premature birth during pregnancy and can be fatal for those with weakened immune systems. Antibiotics are used for treatment.

A major “diablo wind”—notorious in autumn for its hot, dry gusts—is expected to whip up across Northern California on Thursday evening, causing humidity levels to drop and raising the risk of wildfires.

Forecasters have issued red flag warnings for fire danger until Saturday from the central coast through the San Francisco Bay Area and into northern Shasta County, not far from the Oregon border.

Sustained winds reaching 35 mph (56 kph) are expected in many areas, with possible gusts topping 65 mph (104 kph) along mountaintops, according to the National Weather Service.

“This could end up being the most significant wind event for this year so far,” said meteorologist Brayden Murdock with the service’s Bay Area office. “We want to tell people to be cautious.”

During a diablo wind, common in the fall, the air is so dry that relative humidity levels plunge, drying out vegetation and making it ready to burn. The name—”diablo” is Spanish for “devil”—is informally applied to a hot wind that blows near the San Francisco region from the interior toward the coast as high pressure builds over the West.

Pacific Gas and Electric Co. said it was prepared to turn off power to a small number of customers in areas where strong gusts could damage electrical equipment and spark blazes.

Targeted power shutoffs were also possible in Southern California, where another notorious weather phenomenon, the Santa Ana winds, are expected Friday and Saturday.

Winds around greater Los Angeles won’t be as powerful as up north, with gusts between 25 and 40 mph (40 and 64 kph) possible in mountains and foothills, said Mike Wofford, a meteorologist with the weather service’s Los Angeles-area office.

“I think it’s going to be more moderate,” he said Wednesday. “But the risk of fires is still there.”

But snow is in the forecast for the mountaintops around Lake Tahoe, where up to 2 inches (5 centimeters) was forecast by Friday morning, according to the National Weather Service in Reno, Nevada. Winds around Lake Tahoe could gust up to 70 mph (113 kph).

The service also issued its first freeze warning of the season along the Sierra’s eastern front effective from 2 a.m. to 9 a.m. Friday from south of Carson City to the north through Reno into Lassen, Sierra and Plumas counties in California where temperatures could dip into the low 20s Fahrenheit (-5 Celsius).

“Frost and freeze conditions could kill crops, other sensitive vegetation and possibly damage unprotected outdoor plumbing,” the service said.

Samsung Electronics has postponed taking deliveries of ASML chipmaking equipment for its upcoming factory in Texas as it has yet to win any major customers for the project, three people familiar with the matter said.

Samsung has been also holding off on placing orders to some other suppliers for the $17 billion factory in Taylor city, prompting them to look for other customers and send staff deployed on site back home, three other people familiar with the matter said.

The delay in equipment deliveries is a fresh setback to the Taylor project, which is at the heart of Samsung chairman Jay Y. Lee’s ambition to expand beyond its bread-and-butter memory chips into contract chip manufacturing, which Taiwan’s TSMC dominates.

It underscores the widening gulf between Samsung and rivals such as TSMC and SK Hynix which are ramping up production of high-end chips to cater to booming demand from artificial intelligence applications.

ASML, the world’s biggest chipmaking equipment supplier, cut its 2025 sales forecast on Tuesday, citing weakness in markets other than AI, and delayed fabs.

The Dutch company did not name clients who have delayed their fabs. Reuters is the first to report that Samsung has pushed back deliveries of some ASML equipment.

Two of the sources said the delayed shipments to Samsung’s Taylor factory involve ASML’s advanced chipmaking equipment called extreme ultraviolet (EUV) lithography.

One of them said the deliveries were scheduled earlier this year but the machines have not been shipped yet. The third source said Samsung has pushed back delivery of some ASML equipment to the factory, without elaborating on the equipment or the revised delivery schedule.

EUV machines, which cost around $200 million each, create design features on silicon wafers by using beams of light and are widely used to manufacture advanced chips found in smartphones, electronic devices and AI servers.

It was not clear how many EUV machines Samsung had ordered or what payment terms it had entered into.

ASML and Samsung declined to comment on the ASML matter. All of the sources Reuters spoke to declined to be identified as they were not authorised to talk to the media.

‘Stranded asset’

Samsung said in April production at the Taylor plant would begin in 2026 instead of 2024. Samsung’s Lee told Reuters earlier this month that the company was facing challenges on the factory.

Sources and analysts said there was a risk of further delays.

“Without new volume clients, even the 2026 timetable looks challenging… We see a possibility of a further delay and an asset write-off,” Macquarie analysts said in a September report, adding the fab could be “a stranded asset”.

Lee Min-hee, an analyst at BNK Investment & Securities, said that if Samsung does not place orders for other equipment by early next year, it could signal further delays, given the lead time required to start production.

The South Korean firm aims to complete construction of the building by early next year, a person familiar with the matter said.

Samsung said in a statement to Reuters that there is no change to its plan to start production of its Taylor fab in 2026, and the return of its personnel is part of a routine rotational shift.

Widening gap with TSMC

Despite years of efforts to take on TSMC, Samsung’s market share in contract manufacturing, done in factories known as fabs or foundries in the semiconductor industry, dropped 8 percentage points in the last five years to 11% as of the first quarter of 2024, while TSMC’s market share rose to 61.7% over the same period, according to data from research firm Statista.

Samsung’s market share erosion underscores technological challenges the firm faces in mastering advanced chip manufacturing to lure the likes of Apple and Nvidia away from TSMC, analysts said.

ASML finance chief Roger Dassen said on Tuesday that there are “very specific competitive issues in the foundry business”, and some customers are slowly ramping up production of advanced chips and delaying fabs.

Analysts say Intel, which is suffering one of its worst periods and cut its capital spending plan for 2025, is also partly the reason for ASML’s weaker outlook.

By contrast, TSMC expected on Thursday its first fab in Arizona to see volume production in 2025 and said the factory has received strong commitment from U.S. customers.

Samsung’s struggles in its foundry business has also impacted its factories in South Korea, where it is grappling with low production yields of its most advanced 3-nanometer chips, people familiar with the matter said. Samsung is also delaying investments into new foundry chip lines in the city of Pyeongtaek, south of Seoul, two of the people said.

Samsung declined to comment on the Korean factory matter.

Reflecting Samsung’s slowing capacity expansion, ASML reported that sales to South Korea—home to Samsung and its smaller rival SK Hynix—shrunk by one-third in the third quarter to 889 million euros ($965 million) from the previous quarter, according to its presentation materials.

Samsung is also losing ground in its core memory chip market, with SK Hynix outshining it to become the most dominant supplier of pricey high-bandwidth memory (HBM) chips used to build Nvidia’s AI chipsets.

“I think today without AI, the market would be very sad if you ask me,” ASML CEO Christophe Fouquet said in a conference call this week, adding a slower-than-expected recovery for mobile devices and PCs will extend into next year.

($1 = 0.9213 euros)

—Heekyong Yang, Hyunjoo Jin and Toby Sterling, Reuters

Just weeks before the U.S. presidential election, TikTok approved advertisements that contained election disinformation even though it has a ban on political ads, according to a report published Thursday by the nonprofit Global Witness.

The technology and environmental watchdog group submitted ads that it designed to test how well systems at social media companies work in detecting different types of election misinformation.

The group, which did a similar investigation two years ago, did find that the companies—especially Facebook—have improved their content-moderation systems since then.

But it called out TikTok for approving four of the eight ads submitted for review that contained falsehoods about the election. That’s despite the platform’s ban on all political ads in place since 2019.

The ads never appeared on TikTok because Global Witness pulled them before they went online.

“Four ads were incorrectly approved during the first stage of moderation, but did not run on our platform,” TikTok spokesman Ben Rathe said. “We do not allow political advertising and will continue to enforce this policy on an ongoing basis.”

Facebook, which is owned by Meta Platforms Inc., “did much better” and approved just one of the eight submitted ads, according to the report.

In a statement, Meta said while “this report is extremely limited in scope and as a result not reflective of how we enforce our policies at scale, we nonetheless are continually evaluating and improving our enforcement efforts.”

Google’s YouTube did the best, Global Witness said, approving four ads but not letting any publish. It asked for more identification from the Global Witness testers before it would publish them and “paused” their account when they didn’t. However, the report said it is not clear whether the ads would have gone through had Global Witness provided the required identification.

Google did not immediately respond to a message for comment.

Companies nearly always have stricter policies for paid ads than they do for regular posts from users. The ads submitted by Global Witness included outright false claims about the election—such as stating that Americans can vote online—as well as false information designed to suppress voting, like claims that voters must pass an English test before casting a ballot. Other fake ads encouraged violence or threatened electoral workers and processes.

The ads Global Witness submitted were text-based, but the group said it translated them into what it called “algospeak.” This is a widely used trick to try to bypass internet companies’ text-focused content moderation systems by substituting numbers and symbols as stand-in for letters, making it harder for automated systems to “read” the text.

—Barbara Ortutay, AP Technology Writer