The Guardian notes that "While leap seconds pass by unnoticed for most people, they can cause problems for a range of systems that require an exact, uninterrupted flow of time, such as satellite navigation, software, telecommunication, trade and even space travel."

So now Nature magazine reports that "The practice of adding 'leap seconds' to official clocks to keep them in sync with Earth's rotation will be put on hold from 2035, the world's foremost metrology body has decided." The decision was made by representatives from governments worldwide at the General Conference on Weights and Measures (CGPM) outside Paris on 18 November. It means that from 2035, or possibly earlier, astronomical time (known as UT1) will be allowed to diverge by more than one second from coordinated universal time (UTC), which is based on the steady tick of atomic clocks. Since 1972, whenever the two time systems have drifted apart by more than 0.9 seconds, a leap second has been added....

Facebook's parent company, Meta, and Google are among the tech companies that have called for leap seconds to be scrapped. The CGPM — which also oversees the international system of units (SI) — has proposed that no leap second should be added for at least a century, allowing UT1 and UTC to slide out of sync by about 1 minute. But it plans to consult with other international organizations and decide by 2026 on what upper limit, if any, to put on how much they be allowed to diverge....

Although in the long term Earth's rotation slows due to the pull of the Moon, a speed-up since 2020 has also made the issue more pressing, because for the first time, a leap second might need to be removed, rather than added. UTC has only ever had to slow a beat to wait for Earth, not skip ahead to catch up with it. "It's kind of being described as a Y2K issue, because it's just something that we've never had to deal with," said Elizabeth Donley, who leads the Time and Frequency division at the National Institute of Standards and Technology, in Boulder, Colorado.

The Justice Department has opened an antitrust investigation into the owner of Ticketmaster, whose sale of Taylor Swift concert tickets descended into chaos this week, The New York Times reported Friday, citing sources. The investigation is focused on whether Live Nation Entertainment has abused its power over the multibillion-dollar live music industry. From the report: That power has been in the spotlight after Ticketmaster's systems crashed while Ms. Swift fans were trying to buy tickets in a presale for her upcoming tour, but the investigation predates the botched sale, the people said. Staff members at the agency's antitrust division have in recent months contacted music venues and players in the ticket market, asking about Live Nation's practices and the wider dynamics of the industry, said the people, who spoke on condition of anonymity because the investigation is sensitive. The inquiry appears to be broad, looking at whether the company maintains a monopoly over the industry, said one of the people. Officials in the Biden administration have spent the last two years trying to push the boundaries of antitrust law. The Justice Department has mounted several challenges to major mergers, successfully convincing a judge to block Penguin Random House's purchase of Simon & Schuster but losing some other cases. The Federal Trade Commission has sued to block Meta, Facebook's parent company, from acquiring a small virtual reality start-up.

On November 15 Meta unveiled a new large language model called Galactica, designed to assist scientists. But instead of landing with the big bang Meta hoped for, Galactica has died with a whimper after three days of intense criticism. Yesterday the company took down the public demo that it had encouraged everyone to try out. From a report: Meta's misstep -- and its hubris -- show once again that Big Tech has a blind spot about the severe limitations of large language models. There is a large body of research that highlights the flaws of this technology, including its tendencies to reproduce prejudice and assert falsehoods as facts.

Galactica is a large language model for science, trained on 48 million examples of scientific articles, websites, textbooks, lecture notes, and encyclopedias. Meta promoted its model as a shortcut for researchers and students. In the company's words, Galactica "can summarize academic papers, solve math problems, generate Wiki articles, write scientific code, annotate molecules and proteins, and more." But the shiny veneer wore through fast. Like all language models, Galactica is a mindless bot that cannot tell fact from fiction. Within hours, scientists were sharing its biased and incorrect results on social media.

Meta has confirmed that it's removing addresses, "interested in", political views and religion from Facebook profiles as of December 1st. Engadget reports: The move is meant to make Facebook "easier to navigate and use," a spokesperson told TechCrunch. If you've filled out any of these fields, you'll get a notification about the change. Other details you provide, such as your contact information and relationship status, will persist. You can download a copy of your Facebook data before December 1st if you're determined to preserve it, and you still have control over who can see the remaining profile content.

Meta has fired or disciplined more than two dozen employees and contractors over the last year whom it accused of improperly taking over user accounts, in some cases allegedly for bribes, The Wall Street Journal reported Thursday, citing people familiar with the matter and documents. From the report: Some of those fired were contractors who worked as security guards stationed at Meta facilities and were given access to the Facebook parent's internal mechanism for employees to help users having trouble with their accounts, according to the documents and people familiar with the matter. The mechanism, known internally as "Oops," has existed since Facebook's early years as a means for employees to help users they know who have forgotten their passwords or emails, or had their accounts taken over by hackers.

As part of the alleged abuse of the system, Meta says that in some cases workers accepted thousands of dollars in bribes from outside hackers to access user accounts, the people and documents say. The disciplinary actions are part of a lengthy internal probe led by Meta executives, according to the documents and one of the people. "Individuals selling fraudulent services are always targeting online platforms, including ours, and adapting their tactics in response to the detection methods that are commonly used across the industry," said Meta spokesman Andy Stone. He added that the company "will keep taking appropriate action against those involved in these kinds of schemes."

Meta employees are taking aim at Mark Zuckerberg in employee reviews on Blind, the anonymous forum. From a report: Some reviews, posted on Wednesday -- the day Meta laid off 13% of its workforce -- are negative, although others are more positive. One user likened the layoffs to the "hunger games" and another said the Facebook owner had an "uncertain future." Insider surveyed the workplace community app, where staff can air their grievances in posts and reviews, to see what was being said about Meta and its CEO. Some 44 employee reviews of Meta were posted on Blind on Wednesday and Thursday this week.

"The Metaverse will be our slow death," one user, who called themselves a senior software developer, posted on Wednesday. They added: "Mark Zuckerberg will single-handedly kill a company with the meta-verse." Zuckerberg apologized to staff for the need to cut 11,000 jobs, admitting that he "got this wrong". Blind users must provide their work email email address, job title and employer when joining the platform so the company can "gauge the professional status" of posters, according to its website. A user's employment is not officially verified, however. Blind said it occasionally sent prompts to users to "re-verify" their accounts. Rick Chen, head of public relations at Blind, told Insider: "Nearly all of the reviews posted have been written by current employees of the respective companies at the time of writing, as people generally cannot access Blind after they are laid off or resign."

Meta is exiting its Portal smart display business and will wind down work on smartwatch projects, reports Reuters citing company executives during an employee townhall meeting on Friday. The news comes two days after Meta announced about 11,000 job cuts, or 13% of its workforce, "the first mass layoffs in the company's 18-year history."

Amazon Chief Executive Andy Jassy is leading a cost-cutting review of the tech giant and paring back on businesses at the company that haven't been profitable,
Wall Street Journal reported Thursday, citing people familiar with the matter. From the report: The Seattle-based company, whose stock is down about 45% year to date, has been experiencing a slowdown in its core retail business as it tries to manage costs from its logistics network. Other technology companies have been making cuts to better navigate a potential recessionary environment. This week, Facebook parent Meta Platforms said it would cut more than 11,000 workers, or 13% of staff.

As part of the monthslong cost-cutting review, Amazon has told employees in certain unprofitable divisions to look for jobs elsewhere in the company, because the teams they were working on were being suspended or closed, some of the people said. Efforts to scrutinize expenses across a sprawling array of businesses have become common at the world's largest technology companies. Alphabet's Google has slowed the pace of hiring and scaled back support for a startup incubator, and Chief Executive Sundar Pichai has voiced concerns this year about employee productivity.

Mark Zuckerberg, in a blog post: Today I'm sharing some of the most difficult changes we've made in Meta's history. I've decided to reduce the size of our team by about 13% and let more than 11,000 of our talented employees go. We are also taking a number of additional steps to become a leaner and more efficient company by cutting discretionary spending and extending our hiring freeze through Q1. I want to take accountability for these decisions and for how we got here. I know this is tough for everyone, and I'm especially sorry to those impacted.

At the start of Covid, the world rapidly moved online and the surge of e-commerce led to outsized revenue growth. Many people predicted this would be a permanent acceleration that would continue even after the pandemic ended. I did too, so I made the decision to significantly increase our investments. Unfortunately, this did not play out the way I expected. Not only has online commerce returned to prior trends, but the macroeconomic downturn, increased competition, and ads signal loss have caused our revenue to be much lower than I'd expected. I got this wrong, and I take responsibility for that.

In this new environment, we need to become more capital efficient. We've shifted more of our resources onto a smaller number of high priority growth areas -- like our AI discovery engine, our ads and business platforms, and our long-term vision for the metaverse. We've cut costs across our business, including scaling back budgets, reducing perks, and shrinking our real estate footprint. We're restructuring teams to increase our efficiency. But these measures alone won't bring our expenses in line with our revenue growth, so I've also made the hard decision to let people go.

An anonymous reader quotes a report from the New York Times: The user on Gab who identifies as Nora Berka resurfaced in August after a yearlong silence on the social media platform, reposting a handful of messages with sharply conservative political themes before writing a stream of original vitriol. The posts mostly denigrated President Biden and other prominent Democrats, sometimes obscenely. They also lamented the use of taxpayer dollars to supportUkraine in its war against invading Russian forces, depicting Ukraine's president as a caricature straight out of Russian propaganda. The fusion of political concerns was no coincidence. The account was previously linked to the same secretive Russian agency that interfered in the 2016 presidential election and again in 2020, the Internet Research Agency in St. Petersburg, according to the cybersecurity group Recorded Future. It is part of what the group and other researchers have identified as a new, though more narrowly targeted, Russian effort ahead ofTuesday's midterm elections. The goal, as before, is to stoke anger among conservative voters and to undermine trust in the American electoral system. This time, it also appears intended to undermine the Biden administration's extensive military assistance to Ukraine.

"It's clear they are trying to get them to cut off aid and money to Ukraine," said Alex Plitsas, a former Army soldier and Pentagon information operations official now with Providence Consulting Group, a business technology company. The campaign -- using accounts that pose as enraged Americans like Nora Berka -- have added fuel to the most divisive political and cultural issues in the country today. It has specifically targeted Democratic candidates in the most contested races, including the Senate seats up for grabs in Ohio, Arizona and Pennsylvania, calculating that a Republican majority in the Senate and the House of Representatives could help the Russian war effort. The campaigns show not only how vulnerable the American political system remains to foreign manipulation but also how purveyors of disinformation have evolved and adapted to efforts by the major social media platforms to remove or play down false or deceptive content. The agencies urged people not to like, discuss or share posts online from unknown or distrustful sources. They did not identify specific efforts, but social media platforms and researchers who track disinformation have recently uncovered a variety of campaigns by Russia, China and Iran.

These are much smaller campaigns than those in the 2016 election, where inauthentic accounts reached millions of voters across the political spectrum on Facebook and other major platforms. The efforts are no less pernicious, though, in reaching impressionable users who can help accomplish Russian objectives, researchers said. "The audiences are much, much smaller than on your other traditional social media networks," said Brian Liston, a senior intelligence analyst with Recorded Future who identified the Nora Berka account. "But you can engage the audiences in much more targeted influence ops because those who are on these platforms are generally U.S. conservatives who are maybe more accepting of conspiratorial claims." Some characteristics of an inauthentic user to look out for include: no profile picture, no identifying biographical details, and posts exclusively on political issues that often include false or misleading posts and little engagement. They may also link to obscure websites like electiontruth.net, which Recorded Future said was almost certainly linked to the Russian campaign.

A juvenile bar-tailed godwit -- known only by its satellite tag number 234684 -- has flown 13,560 kilometres from Alaska to the Australian state of Tasmania without stopping, appearing to set a new world record for marathon bird flights. From a report: The five-month-old bird set off from Alaska on 13 October and satellite data appeared to show it did not stop during its marathon flight which took 11 days and one hour. Tagged in Alaska, the bar-tailed godwit, Limosa lapponica, flew at least 13,560km (8,435 miles) before touching down at Ansons Bay in north-east Tasmania.

The previous record was held by an adult male of the same species -- 4BBRW -- that flew 13,000km (8,100 miles) last year, beating his own previous record of 12,000km the year before. According to a Facebook post from the Pukorokoro Miranda Shorebird Centre in New Zealand, 4BBRW's record had been "blown out of the water by this young upstart." Scientists track the bird using a 5G satellite tag attached to its lower back. According to data from the Max Planck Institute for Ornithology's bird tracking project, the migratory bird took a route to the west of Hawaii, continuing over open ocean and flying over the Pacific island nation of Kiribati on 19 October.

Meta is planning to begin large-scale layoffs this week, WSJ reported over the weekend, citing people familiar with the matter, in what could be the largest round in a recent spate of tech job cuts after the industry's rapid growth during the pandemic. From the report: The layoffs are expected to affect many thousands of employees and an announcement is planned to come as soon as Wednesday, according to the people. Meta reported more than 87,000 employees at the end of September. Company officials already told employees to cancel nonessential travel beginning this week, the people said.

The planned layoffs would be the first broad head-count reductions to occur in the company's 18-year history. While smaller on a percentage basis than the cuts at Twitter Inc. this past week, which hit about half of that company's staff, the number of Meta employees expected to lose their jobs could be the largest to date at a major technology corporation in a year that has seen a tech-industry retrenchment.

In 2004 NBC's news show "Dateline" began airing "To Catch a Predator" segments, in which a vigilante group posed online as minors to lure sex predators into in-person meetings — where they were then arrested by police.

The New Yorker looks at its cultural impact: Although there were only twenty episodes of the series, in three years, it's "this touchstone that I grew up with and that millions of people grew up with," Paul Renfro, a professor of history at Florida State University and the author of "Stranger Danger: Family Values, Childhood, and the American Carceral State," said. "It shaped how people think about sexual violence in ways that we haven't fully grappled with." The show focussed on the threat from strangers on the Internet, even though most victims of child sexual abuse are harmed by someone known to them. "On the show, it's not the family, it's not priests or rabbis or other authority figures who pose a threat to children, it's this devious stranger," Renfro said. The show's influence helped spur the passage of the Adam Walsh Act, in 2006, which created publicly searchable databases of people convicted of certain sex crimes. (There's little evidence that sex-offender registries have been effective at reducing sexual offenses.)
But today, "amateur predator hunting has come back into style," the article notes, citing the proliferation of online groups. "Recently, the Washington Post found more than a hundred and sixty, which have been responsible for nearly a thousand stings this year."

And then the New Yorker interviewed a woman named Cam, who with her husband and her brother-in-law decided to form "the Permian Basin Predator Patrol" — broadcasting their sting operations and humiliations of potential perpetrators on YouTube: [S]oon after the channel started drawing attention, they were called to a meeting at the Odessa Police Department. According to Cam, officers made it clear that they disapproved of their activities. "We were told we can't be involved with them, and that we can't send them anything directly," she said. "One, we're endangering ourselves, and, two, we're giving them more work — that's what it seemed like they were saying."

"We are very mindful of not trying to entrap a suspect," Lieutenant Brad Cline, who works in the Odessa Police Department's Crimes Against Persons Unit, said. "Taking a predator into custody can be very dangerous as well."
The article points out that "To Catch a Predator" was cancelled when Texas man Bill Conradt decided not to follow-up on his online messages — but "When a SWAT team burst into his house, trailed by a camera crew, Conradt shot himself."

So what did Cam's group do when the Odessa Police Department declined their help? The Permian Basin Predator Patrol continued to make videos. If she couldn't contribute to an arrest, Cam thought, at least she could get the word out to the public. She became an expert at figuring out the identities of the men she was chatting with, even when they used fake names.... Sometimes she'd find a man's family on Facebook and send his mother screenshots of the obscene messages he'd sent, or call his employer. "I believe three of them have been let go from their jobs," she said.

A sting by the Predator Catchers Indianapolis led to a man's conviction for child solicitation.... Although YouTube's predator hunters tend to portray themselves as the unequivocal good guys (Cam is an exception — most are men), their track record is more mixed.... The Ohio-based group Dads Against Predators has reportedly been banned from local grocery stores for causing disturbances. In 2018, a twenty-year-old in Connecticut hanged himself after a confrontation with a predator-hunter group. One video by the Permian Basin Predator Patrol ends with a man weeping, then running into traffic. (Cam said that she asked police to perform a welfare check on him, but she's not sure if it occurred.)

By Tuesday morning Mastodon had gained 123,562 new users since October 27 (the site told TechCrunch) and had 528,607 active users. But by Saturday the number of new users had nearly doubled, to 230,000, reports CNN — with 655,000 active users.

In fact, for every 363 active users on Twitter, there's now one on Mastodon, CNN's figures suggest (since Twitter has nearly "238 million daily active monetizable users"). Exploring the recent spike, they note that Mastodon "has a similar look to Twitter, with a timeline of short updates sorted chronologically rather than algorithmically. It lets users join a slew of different servers run by various groups and individuals, rather than one central platform controlled by a single company like Twitter, Instagram, or Facebook." Unlike larger social networks, Mastodon is both free to use and free of ads. It's operated by a nonprofit run by Mastodon creator Eugen Rochko, and is supported via crowdfunding... "It is not as large as Twitter, obviously, but it is the biggest that this network has ever been," said Rochko, who originally created Mastodon as more of a project than a consumer product (and, yes, its name was inspired by the heavy metal band Mastodon)....

A lot of Mastodon's features and layout (particularly in its iOS app) will look and feel familiar to current Twitter users, though with some slightly different verbiage; you can follow others, create short posts (there's a 500 character limit, and you can upload images and videos), favorite or repost other users' posts, and so on.... There are some key differences, particularly in how the network is set up. Because Mastodon users' accounts are hosted on a slew of different servers, the costs of hosting users is spread among many different people and groups. But that also means users are spread out all over the place, and people you know can be hard to find.
CNN also notes the problem with signing up for a Mastodon server: "some of which are open to anyone, some of which require an invitation (you can also run your own server). There is a server operated by the nonprofit behind Mastodon, Mastodon.social, but it's not accepting more users."

Meta's Instagram will soon allow artists to create and sell their own NFTs both right on the social media platform and off it. Axios reports: IG's feature will roll out to a small group of select creators in the U.S. to start, according to Meta. The first creators tapped to test the feature include photographer Isaac âoeDriftâ Wright, known as DrifterShoots, and artist Amber Vittoria. Meta won't charge fees for posting or sharing an NFT on IG, though, app store fees still apply. Separately, there will be a "professional mode" for Facebook profiles for creators to build a social media presence separate from their personal one. Artist royalties appear to be a part of the plan.

Minting or the creating of NFTs on IG will start on Polygon, a boon for the layer-2 blockchain (a separate blockchain built on top of Ethereum) given the potential onboarding of IG's billion active users. The price of Polygon's token MATIC jumped 17% from Wednesday evening to Thursday morning, boosted by IG news but also, because JPMorgan conducted its first live DeFi trade using that blockchain. The platform is adding support for Solana blockchain and Phantom wallet with the latest feature update, adding them to the list of already-supported wallets such as MetaMask, Coinbase Wallet, Dapper Wallet, Rainbow and Trust Wallet. Ethereum and Flow blockchains are already supported. Info for selected collections with OpenSea metadata, like collection name and description, will show up on IG.

An anonymous reader quotes a report from Reason Magazine: The Babylon Bee this week joined The Onion in urging the Supreme Court to defend the First Amendment against an Ohio law that makes parody a felony. The case, which the Institute for Justice is asking the Court to take up, involves Parma resident Anthony Novak, who in 2016 was prosecuted for violating a state law against using a computer to "disrupt, interrupt, or impair the functions of any police, fire, educational, commercial, or governmental operations." Novak supposedly did that by creating a parody of the Parma Police Department's Facebook page. [...]

For obvious reasons, the right-leaning Bee, like the left-leaning Onion, is alarmed by the implication that people have no recourse against cops who arrest them for making fun of government agencies. "The Bee is serving a brutal life sentence in Twitter jail as we speak," says its amicus brief (PDF) in Novak v. City of Parma. "Its writers would very much like to avoid a consecutive sentence in a government-run facility." The premise of Novak's prosecution was that he had disrupted police operations by prompting calls about his parody to the department's nonemergency line. "Left in the hands of the Sixth Circuit and the Parma PD (and other like-minded law enforcement), the speech-stifling Ohio statute used to go after Mr. Novak empowers state officials to search, arrest, jail, and prosecute parodists without fear of ever being held accountable," the Bee says. "The upshot for The Bee is that, in Ohio at least, its writers could be jailed for many, if not most, of the articles The Bee publishes, provided that someone contacted law enforcement -- or another entity 'protected' by [Ohio's law] -- to tell them that the articles exist."

Consider the March 3 Bee story headlined "Donut Sales Surge as Police Departments Re-Funded." If someone "had called the Parma Police Department to let them know that The Bee had published the article," the brief suggests, the publication "could have been charged with a felony, its offices searched, and its writers arrested and jailed for days, all without consequence for the parties doing the charging, arresting, jailing, and searching." Likewise if an officer's "passive-aggressive brother-in-law had forwarded the article" to the cop's official email address, thereby "interrupt[ing]" his work. Given the broad wording of Ohio's law, which refers to "governmental operations" generally, Bee articles about federal agencies, such as its August 12 report on the FBI's search of Mar-a-Lago, also could be treated as grounds for arrest. "Had a caller contacted the FBI field office in Cleveland or Cincinnati" to "express outrage over the suspicious timing of the FBI's raid on Melania Trump's Mar-a-Lago closet and Attorney General Garland's acquisition of a haute couture wardrobe," the Bee notes, that could be the basis for a felony charge in Ohio. On the First Amendment issues raised by this case, both The Onion and The Babylon Bee see eye to eye.

"The Onion may be staffed by socialist wackos, but in their brief defending parody to this Court, they hit it out of the park," the Bee says. "Parody has a unique capacity to speak truth to power and to cut its subjects down to size. Its continued protection under the First Amendment is crucial to preserving the right of citizens to effectively criticize the government."

The macro story unfolding today is all the layoffs taking place in the tech industry. "Tech giants including Meta and Amazon have been slowing down their hiring for months, while smaller tech companies such as Robinhood and Coinbase have announced layoffs," reports the New York Times. "But rarely have so many job cuts and hiring freezes in the industry been disclosed on the same day." From the report: The technology industry's slowdown came into even sharper relief on Thursday as Amazon publicly said it had paused hiring for its corporate work force and several other technology companies announced job cuts. [...] At the same time, Lyft said it would cut 13 percent of its employees, or about 650 of its 5,000 workers. Stripe, a payment processing platform, said it would cut 14 percent of its employees, roughly 1,100 jobs. [...] Tech companies have led the way for the U.S. economy over the past decade, lifting the stock market during the worst days of the coronavirus pandemic. But in recent weeks, many of the largest firms reported financial results that suggested they were feeling the impact of global economic jitters, soaring inflation and rising interest rates.

Social media companies in particular have been grappling with a pullback in digital advertising over the last few months. Meta, which owns Facebook and Instagram, said last week that its head count would remain "roughly flat" through the end of next year. The company plans to shrink some teams and hire only for high-priority areas. Snap, Snapchat's parent company, laid off 20 percent of its employees in August, blaming challenging macroeconomic conditions. Last week, Microsoft told investors that new hires in this quarter "should be minimal." Alphabet, which owns Google and YouTube, also said that in this quarter it would hire fewer than half the number of people it added in the third quarter.

More layoffs at tech companies are in the works. Elon Musk, who bought Twitter for $44 billion last week, has ordered cuts across the company, which employs about 7,500 people. Workers at Twitter have started circulating a "Layoff Guide" with tips on how to handle being laid off. On Thursday, Lyft said it had decided on layoffs in the face of "a probable recession sometime in the next year." All teams will be affected, said Logan Green and John Zimmer, the company's founders, in an email to employees. Over the summer, Lyft cut 2 percent of its employees, mostly as a result of shutting down its car rental business, and froze hiring. But the company still has "to become leaner," its founders said. It is "not immune to the realities of inflation and a slowing economy," which have led to increasing ride-share insurance costs. Lyft also said it planned to sell its first-party vehicle service business and expected employees on that team to be offered jobs at the acquiring company.

Last week, Meta announced an AI-powered audio compression method called "EnCodec" that can reportedly compress audio 10 times smaller than the MP3 format at 64kbps with no loss in quality. Meta says this technique could dramatically improve the sound quality of speech on low-bandwidth connections, such as phone calls in areas with spotty service. The technique also works for music. Ars Technica reports: Meta debuted the technology on October 25 in a paper titled "High Fidelity Neural Audio Compression," authored by Meta AI researchers Alexandre Defossez, Jade Copet, Gabriel Synnaeve, and Yossi Adi. Meta also summarized the research on its blog devoted to EnCodec.

Meta describes its method as a three-part system trained to compress audio to a desired target size. First, the encoder transforms uncompressed data into a lower frame rate "latent space" representation. The "quantizer" then compresses the representation to the target size while keeping track of the most important information that will later be used to rebuild the original signal. (This compressed signal is what gets sent through a network or saved to disk.) Finally, the decoder turns the compressed data back into audio in real time using a neural network on a single CPU.

Meta's use of discriminators proves key to creating a method for compressing the audio as much as possible without losing key elements of a signal that make it distinctive and recognizable: "The key to lossy compression is to identify changes that will not be perceivable by humans, as perfect reconstruction is impossible at low bit rates. To do so, we use discriminators to improve the perceptual quality of the generated samples. This creates a cat-and-mouse game where the discriminator's job is to differentiate between real samples and reconstructed samples. The compression model attempts to generate samples to fool the discriminators by pushing the reconstructed samples to be more perceptually similar to the original samples."

An anonymous reader quotes a report from The Intercept: The Department of Homeland Security is quietly broadening its efforts to curb speech it considers dangerous, an investigation by The Intercept has found. Years of internal DHS memos, emails, and documents -- obtained via leaks and an ongoing lawsuit, as well as public documents -- illustrate an expansive effort by the agency to influence tech platforms. The work, much of which remains unknown to the American public, came into clearer view earlier this year when DHS announced a new "Disinformation Governance Board": a panel designed to police misinformation (false information spread unintentionally), disinformation (false information spread intentionally), and malinformation (factual information shared, typically out of context, with harmful intent) that allegedly threatens U.S. interests. While the board was widely ridiculed, immediately scaled back, and then shut down within a few months, other initiatives are underway as DHS pivots to monitoring social media now that its original mandate -- the war on terror -- has been wound down.

Behind closed doors, and through pressure on private platforms, the U.S. government has used its power to try to shape online discourse. According to meeting minutes and other records appended to a lawsuit filed by Missouri Attorney General Eric Schmitt, a Republican who is also running for Senate, discussions have ranged from the scale and scope of government intervention in online discourse to the mechanics of streamlining takedown requests for false or intentionally misleading information. [...] There is also a formalized process for government officials to directly flag content on Facebook or Instagram and request that it be throttled or suppressed through a special Facebook portal that requires a government or law enforcement email to use. At the time of writing, the "content request system" at facebook.com/xtakedowns/login is still live. These are the key takeaways from the report: - Though DHS shuttered its controversial Disinformation Governance Board, a strategic document reveals the underlying work is ongoing.
- DHS plans to target inaccurate information on 'the origins of the COVID-19 pandemic and the efficacy of COVID-19 vaccines, racial justice, U.S. withdrawal from Afghanistan, and the nature of U.S. support to Ukraine."
- Facebook created a special portal for DHS and government partners to report disinformation directly.
- The work is primarily done by CISA, a DHS sub-agency tasked with protecting critical national infrastructure.
- DHS, the FBI, and several media entities are having biweekly meetings as recently as August.
- DHS considered countering disinformation relating to content that undermines trust in financial systems and courts.
- The FBI agent who primed social media platforms to take down the Hunter Biden laptop story continued to have a role in DHS policy discussions.

Today ProPublica published "the largest-ever analysis of Google's ad practices on non-English-language websites," saying their report shows Google "is funneling revenue to some of the web's most prolific purveyors of false information in Europe, Latin America and Africa," and "reveals how the tech giant makes disinformation profitable...." The company has publicly committed to fighting disinformation around the world, but a ProPublica analysis, the first ever conducted at this scale, documented how Google's sprawling automated digital ad operation placed ads from major brands on global websites that spread false claims on such topics as vaccines, COVID-19, climate change and elections.... The resulting ad revenue is potentially worth millions of dollars to the people and groups running these and other unreliable sites — while also making money for Google.

Platforms such as Facebook have faced stark criticism for failures to crack down on disinformation spread by people and governments on their platforms around the world. But Google hasn't faced the same scrutiny for how its roughly $200 billion in annual ad sales provides essential funding for non-English-language websites that misinform and harm the public. Google's publicly announced policies bar the placement of ads on content that makes unreliable or harmful claims on a range of issues, including health, climate, elections and democracy. Yet the investigation found Google regularly places ads, including those from major brands, on articles that appear to violate its own policy.

ProPublica's examination showed that ads from Google are more likely to appear on misleading articles and websites that are in languages other than English, and that Google profits from advertising that appears next to false stories on subjects not explicitly addressed in its policy, including crime, politics, and such conspiracy theories as chemtrails. A former Google leader who worked on trust and safety issues acknowledged that the company focuses heavily on English-language enforcement and is weaker across other languages and smaller markets....
The former Google leader suggests Google focuses on English-language problems partly because they're sensitive to bad PR and the posibility of regulatory scrutiny (and because English-language markets have the biggest impact).

Google is spending more money to patrol non-English content, a spokesperson told ProPublica, touting the company's "extensive measures to tackle misinformation... In 2021, we removed ads from more than 1.7 billion publisher pages and 63,000 sites globally. We know that our work is not done, and we will continue to invest in our enforcement systems to better detect unreliable claims and protect users around the world."

But in some cases Google's ads appeared on false online article published years ago, the article points out, "suggesting that the company's failure to block ads on content that appears to violate its rules is a long-standing and ongoing problem... [T]he investigation shows that as one arm of Google helps support fact-checkers, its core ad business provides critical revenue that ensures the publication of falsehoods remains profitable."