After nearly three weeks of escalating rhetoric, The Wire is retracting its reporting on Meta. On Sunday, the nonprofit publication said it had discovered “certain discrepancies” with the material that had informed its reporting on the social media giant since October 6th. “The Wire believes it is appropriate to retract the stories,” the outlet said, pointing to the fact it could not authenticate two emails that were critical to its previous coverage of Meta. One of the emails The Wire said it could not verify includes a message the outlet had attributed to Meta spokesperson Andy Stone.

“Our investigation, which is ongoing, does not as yet allow us to take a conclusive view about the authenticity and bona fides of the sources with whom a member of our reporting team says he has been in touch over an extended period of time,” The Wire said. “We are still reviewing the entire matter, including the possibility that it was deliberately sought to misinform or deceive The Wire.”

Before Sunday’s retraction, The Wire claimed Meta gave Amit Malviya, an information technology official with India’s ruling BJP party, the power to remove posts from Instagram, an assertation Meta has consistently disputed. Rather than backing down after the company shared a comprehensive rebuttal on October 12th, The Wire kept publishing stories that claimed Meta was misleading the public, culminating in an October 15th article that featured a screen recording the outlet claimed showed proof of the original takedown request that kicked off the entire saga. One day later, Meta said an internal investigation found the video showed a Workspace account created on October 13th, suggesting someone made the account to back up The Wire’s reporting.

Meta did not immediately respond to Engadget’s request for comment. Amid all the back and forth, Instagram eventually reinstated the post that prompted The Wire’s investigation in the first place.

The next time you want to download an app on your iPhone, prepare to see more ads. Starting on October 25th, Apple will begin rolling out two additional ad categories to the App Store, according to an email seen by MacRumors. In all countries outside China, the company plans to begin selling ad space within the “Today” tab and at the bottom of individual app listings. Once they arrive, the ads will feature an icon with a blue background to distinguish them from other recommendations.

While Apple has sold search ads through the App Store since 2016, the Today tab was previously reserved for recommendations from the company’s editorial team. “With a Today tab ad, your app can appear prominently on the front page of the App Store — making it some of the first content users see when they begin their App Store visit,” the company states on a support page.

Apple first announced it was bringing ads to the Today tab in July. The expansion is part of a broader push by the company to expand its advertising business. According to a recent report from Bloomberg’s Mark Gurman, the company hopes to nearly triple its advertising revenue to $10 billion annually over the next few years. In addition to more App Store ads, the company reportedly plans to introduce search ads to Maps in the near future.

Snap has introduced a number of features for its Snapchat+ subscription app, including a new "Story Expiration" feature, along with custom notification sounds, camera borders and more. The Story feature looks useful, and appears designed to tempt hardcore users to sign up for the $4 per month service.

With story expiration, subscribers can choose to let their stories expire for lengths of time between one hour and one week. That could allow creators to, say, draw interest with a short story, or let one stay for a longer period to draw the more eyeballs. As it stands now, all stories expire after 24 hours and must be saved or they'll be gone for good. 

Snap

Paid users are also getting access to custom notification sounds, letting them set different tones for different friends so they tell who Snapped them without checking their phone. And the new custom camera color borders option allows users to just what it says, using different hues while capturing content. And finally, Snapchat is bringing custom "exclusive, seasonal and spooky Bitmoji backgrounds" that they can use in their profiles for Halloween. 

Snapchat+ made a strong start, with the company boasting a million subscribers just a couple of months after it debuted. At the time, Snap described it as a service for "passionate" users, and though that's a fraction of its 347 million daily users, it's a significant injection of additional revenue. Snapchat also sells hardware like glasses and recently came out with the Pixy Drone, though Snap had reportedly given up on that product.

When a journalist at The Wire, an independent Indian publication, published a story on October 6th about a meme page’s claim that their Instagram post had been wrongfully removed, it hardly seemed like the kind of story that would draw much attention. The Instagram account, @cringearchivist, was a private account with fewer than 1,000 followers. The fact that their post, a satirical image depicting an Indian government official, was removed for breaking the app’s rules around sexual activity — despite showing nothing of the sort — was odd, but not the kind of thing that might draw international attention.

But in an increasingly bizarre turn of events, the nonprofit newsroom began publishing stories with more explosive claims about what it alleged led up to the removal of @cringearchivist’s post. And, in an even more unusual move, Meta not only refuted the claims but said the publication’s reporting was based on “fabricated” evidence and likely the result of some kind of elaborate hoax.

What happened?

After the initial story on the Instagram takedown, The Wire then started looking more closely at what happened. After not getting a response from Meta, reporters started asking around with sources inside the company. According to what one reporter for The WiretoldNewsLaundry, sources within Meta told them the post had been removed not by Instagram’s moderators but at the behest of Amit Malviya, an official in India’s ruling BJP party who oversees its IT cell.

The Wire then published a followup to its original story on October 10th, with the headline “Exclusive: If BJP’s Amit Malviya Reports Your Post, Instagram Will Take it Down – No Questions Asked.” The story alleged that Malviya had the power to remove Instagram posts thanks to Meta’s controversial cross-check program, which has been credited with shielding high-profile celebrities and politicians from the company’s rules.

It was an explosive allegation. While cross-check has received plenty of scrutiny, no previous reporting had indicated that those privileges might extend to the ability for those outside the company to influence content takedowns.

Meta immediately refuted the story, Meta spokesperson Andy Stone said that cross-check had “nothing to do with” the ability to report posts and that the original Instagram post was removed due to Instagram's automated tools. He also said that “the underlying documentation appears to be fabricated.”

Rather than backing down, though, The Wire published a new story the next day, this one featuring an email – supposedly sent by Stone — in which the comms official blasted staff for allowing the documents to “leak.” But the supposed email only raised more questions about The Wire’s reporting. Most glaringly, the grammar and syntax in the messages was… strange. It used phrases like “for the last one month” and “post which I will tweet about it.” Journalists who cover Meta and frequently interact with Stone pointed out that not only did it not sound like him, it didn’t sound as if it was written by a native English speaker.

Stone also denied sending the email, and again said The Wire seemed to be relying on faked documents. Meta also published its own rebuttal on October 12th. The screenshots, according to the company, were fabricated. The @cringearchivist posts in question were removed by the company’s automated systems, not a human, much less an Indian government official. “We hope that The Wire is the victim of this hoax, not the perpetrator,” the company wrote.

Once again, The Wire responded that it wasn’t backing down. On October 15th, it published yet another story, titled “Meta Said Damaging Internal Email is ‘Fake’, URL 'Not in Use', Here's Evidence They're Wrong.” The lengthy post included several technical explanations about how the alleged emails from Stone were analyzed and verified. It also cited emails from independent security researchers who allegedly backed up their analysis. And, crucially, it included a screen recording from Meta’s Workplace software that allegedly showed proof of the takedown requests.

But, again, The Wire’s supposed evidence only raised new questions about its sources. On October 16th, Meta weighed in again. This time, the company said that an internal investigation revealed the alleged Workplace video was created from a Workplace account created with a free trial of the software on October 13th.

“At this time, we can confirm that the video shared by The Wire that purports to show an internal Instagram system (and which the Wire claims is evidence that their false allegations are true) in fact depicts an externally-created Meta Workplace account that was deliberately set up with Instagram’s name and brand insignia in order to deceive people,” the company wrote. “It is not an internal account. Based on the timing of this account’s creation on October 13, it appears to have been set up specifically in order to manufacture evidence to support the Wire’s inaccurate reporting.”

And, once again, The Wire said it was standing by its reporting. In a statement on October 17th, the publication essentially said it would no longer engage with Meta on the topic. The publication accused the company of attempting to “goad” them into revealing its sources. “We are not prepared to play this game any further,” it said.

Meanwhile, the alleged proof provided by The Wire continued to fall apart. And one of the security researchers who The Wire said had backed up their verification said he had never sent the messages it cited in its reporting.

On October 18th, the publication reversed course, saying it would review its reporting, and would pull the stories from public view while it investigated. “This will include a review of all documents, source material and sources used for our stories on Meta,” The Wire wrote in a statement. “Based on our sources’ consent, we are also exploring the option of sharing original files with trusted and reputed domain experts as part of this process.”

Now what?

At this point, there are still more questions than answers about how and why The Wire’s reporting went so wrong. It’s clear there are serious issues with the “evidence'' it relied on, though it’s not clear whether it intentionally lied or if it was misled as part of some broader scheme. The publication has said numerous times it relied on two separate sources, which suggests the whole thing is more complex than one bad source.

More information is likely to come out in the coming days and weeks as The Wire and others now look more closely at how the story got so out of control. But there’s a reason why the stakes in this particular incident feel so high. India ranks 150 out of 180 in terms of press freedom, according to Reporters Without Borders. And The Wire, a nonprofit publication, is one of a shrinking number of independent newsrooms in the country.

It’s also worth pointing out that some elements of this saga point to real and serious issues about Meta’s policies and how they impact its billions of users. For one, the entire situation began with something many people have experienced: a content moderation decision gone awry due to a mistake in the company’s automated systems. Amid all the back and forth, Instagram ended up reinstating the original Story post from @cringearchivist that had kicked off The Wire’s investigation.

There’s also the fact that Meta has been less than forthcoming about its cross-check rules for celebrities, politicians and other VIPs. Many details we now know about the program only came to light thanks to a company whistleblower and other investigative reporting into the company. The company’s own Oversight Board, which has been working on an advisory opinion for nearly a year about the program, even accused the company of misleading it about the program.

And while there continues to be no evidence that cross-check would enable a company outsider to initiate content takedowns, journalists and activists have long raised questions about whether Meta gives too much leeway to India’s BJP in other policy decisions.

Put all that together and it becomes clear why a publication like The Wire might be so invested in a story like this in the first place. “Our recent coverage of Meta began with an incident that reflected the lack of transparency at the social media giant and its various platforms,” The Wire wrote in its most recent statement. Unfortunately, its own reporting has so far only made things more opaque.

Nearly half of the world’s roughly 7,000 known languages four in ten of them exist without an accompanying written component. These unwritten languages pose a unique problem for modern machine learning translation systems, as they typically need to convert verbal speech to written words before translating to the new language and reverting the text back to speech, but one that Meta has reportedly addressed with its latest open-source language AI advancement.

As part of Meta’s Universal Speech Translator (UST) program which is working to develop real-time speech-to-speech translation so that Metaverse denizens can more easily interact (read: sexually harass one another). As part of this project, Meta researchers looked at Hokkien, an unwritten language spoken throughout Asia’s diaspora and one of Taiwan’s official languages.

Machine learning translation systems typically require extensive labelable examples of the language, both written and spoken, to train on — precisely what unwritten languages like Hokkien don’t have. To get around that, “we used speech-to-unit translation (S2UT) to convert input speech to a sequence of acoustic units directly in the path previously pioneered by Meta,” CEO Mark Zuckerberg explained in a Wednesday blog post. “Then, we generated waveforms from the units. In addition, UnitY was adopted for a two-pass decoding mechanism where the first-pass decoder generates text in a related language (Mandarin), and the second-pass decoder creates units.”

“We leveraged Mandarin as an intermediate language to build pseudo-labels, where we first translated English (or Hokkien) speech to Mandarin text, and we then translated to Hokkien (or English) and added it to training data,” he continued. Currently, the system allows for someone who speaks Hokkien to converse with someone who speaks English, albeit stiltedly. The model can only translate one full sentence at a time but Zuckerberg is confident that the technique can eventually be applied to more languages and will improve to the point of offering real-time translation.

In addition to the models and training data that Meta is already open-sourcing from this project, the company is also releasing a first-of-its-kind speech-to-speech translation benchmarking system based on a Hokkien speech corpus called Taiwanese Across Taiwan, as well as “the speech matrix, a large corpus of speech-to-speech translations mined with Meta’s innovative data mining technique called LASER,” Zuckerberg announced. This system will empower researchers to create speech-to-speech translation (S2ST) systems of their own.

We've seen rollable device concepts from companies like Oppo and TCL, and LG was even working on a commercial rollable smartphone until it quit making mobile devices last year. Now, Lenovo is showing off a laptop with a rollout display, while its mobile division Motorola has a roll-out smartphone — and they look like some of the more practical efforts yet. 

The phone starts out at a very pocketable 4 inches high, but with the click of a button, the OLED panel extends to a normal-sized 6.5 inches. Another click retracts the phone back to its original form. It's usable at both sizes, and the content on the screen adapts to the size, including the home screen, videos and more, as shown below. 

Lenovo also showed off a rollable laptop that starts with a typical landscape display and then rolls up to a square shape, making it better for documents or vertical TikTok style videos. Lenovo VP of design Brian Leonard explained in the video that it's part of Lenovo's research into novel form factors that started with the ThinkPad 360P laptop/tablet. "It can expand into a much larger screen real estate as the content dynamically adjusts to the screen, offering people an unprecedented way of hyper-tasking for productivity, browsing, and more," he explained.

Motorola already makes a folding device, the Razr, that's unfortunately not available in China, but nothing like this. People seem to love these rollable concepts, though — for instance, LG's rollable OLED TV was a popular item at CES 2019. And in some ways, a rolling screen is more elegant than a folding one, as there's no hinge or crease to worry about. While it remains squarely in the concept realm for now, we'd love to see Lenovo take a shot at commercializing it. 

Comcast's recent speed upgrades now extend beyond the high end. The cable giant is increasing speeds for most of its Xfinity tiers across the US. The entry-level Performance Starter (aka Connect) plan is bumping up from 50Mbps to 75Mbps, while the biggest relative gains come to regular Performance (Connect More) customers climbing from 100Mbps to 200Mbps. Performance Pro/Fast service is increasing from 300Mbps to 400Mbps, while Blast/Superfast users will get 800Mbps instead of the previous 600Mbps. And you no longer need Comcast's absolute best plans to cross the gigabit threshold — Extreme Pro and Ultrafast (now Gigabit) customers have upgraded from 900Mbps to 1Gbps.

The flagship Gigabit Extra/x2 plan still peaks at 1.2Gbps for many customers. Comcast is in the midst of deploying 2Gbps service to more states, and in some cases offers 6Gbps access. The telecom hopes to cover over 50 million homes and offices with 2Gbps by the end of 2025, and plans to offer "10G" and next-gen DOCSIS 4.0 service in the future.

These upgrades aren't dramatic in most cases, but they could make a difference at the lower end by enabling higher-quality streaming and better service for multi-person households. The challenge, of course, is that rivals aren't sitting idle. AT&T is deploying 2Gbps and 5Gbps fiber to dozens of urban areas, and Google Fiber will soon provide 8Gbps service on top of restarting expansion. Comcast may be more competitive, but it won't always have the fastest options.

The entry-level increase might also irk regulators. FCC Chairwoman Jessica Rosenworcel recently proposed raising the definition of broadband to 100Mbps. A Performance Starter or Connect customer would fall short of that new goal. If that baseline takes effect, Comcast would have to increase speeds again to satisfy the Commission and help fulfill goals of improving internet access for rural and low-income Americans.

It's been four years since we got the Surface Studio 2, the much-improved followup to Microsoft's ever-so-flexible all-in-one desktop. Surely, the company has something special in store for the next version, right? Well, yes and no. The Surface Studio 2+ is indeed significantly faster than before, thanks to Intel's H35 11th-gen CPU and NVIDIA's RTX 3060 graphics.

But we've also spent most of this year being impressed by Intel's terrific 12th-gen hybrid chips, which deliver vastly better performance than 11th-gen CPUs. (Just imagine how much more intriguing the Studio 2+ would be with a 16-core 12th-gen HX CPU.) You'd think a computer starting at $4,300 would have the fastest hardware available? No wonder it's not called the Surface Studio 3.

Microsoft

If you've been following our coverage of the Studio line, the Studio 2+ will sound pretty familiar. It has the same 28-inch PixelSense screen, a flexible hinge that can be effortlessly pushed down to an easel-like angle, and it packs in all of its hardware in a desktop-friendly base. Following the trend we've seen with all of Microsoft's Surface hardware this year, there aren't any exterior design changes at all. But hey, at least we've finally got Thunderbolt 4 USB-C ports and Dolby Vision HDR.

Microsoft claims the Core i7-11370H CPU in the Surface Studio 2+ is 50 percent faster than the Studio 2, as well as five times faster than the original 2017 modeled. Coupled with the RTX 3060, which is twice as fast as the GTX 1060 in the Studio 2, this new desktop will definitely be a fast performer. But the hardware geek in me can't help but wonder why Microsoft couldn't make Intel's 12th-gen chips work. They're technically more efficient, and it's not as if there wasn't enough time to prepare for new hardware.

Microsoft

I'm sure the Studio 2+ will satisfy general buyers who are willing to pony up big bucks for a uniquely flexible desktop. But I can't think of recommending Intel's 11th-gen chips in any other computer today, unless you're snagging it with a serious discount. For such a premium, they're practically a dealbreaker.

At least the Studio 2+ is well equipped otherwise, shipping with 32GB of DDR4 RAM and a 1TB SSD. (It sure would be nice to see some SSD expansion slots, though). But take note, if you want it to ship with Microsoft's Surface Pen, Keyboard and Mouse, you'll have to spend an extra $300. (That makes it a $4,500 computer with last year's Intel chip! The horror!)

You can pre-order the Surface Studio 2+ today, and it'll be available in select markets on October 25th.

Follow along with the rest of our news from Microsoft's 2022 Surface event.

Blizzard previously admitted that Overwatch 2's launch, which was spoiled by a bunch of bugs, DDoS attacks and other issues, has not met players of the company's expectations. While the company has made a lot of progress to make the game playable — a lot of players couldn't even log in at first — its work is far from done. Now, the developer is trying to make it up to fans by giving out freebies and running events. It will hold several Double Match XP weekends to give players the chance to rack up points and rank up. Blizzard will announce specific dates for the events soon. 

It will also give players who log in from October 25th until the time Season One ends a Cursed Captain Reaper Legendary skin and a Health Pack Weapon Charm. Both items will automatically be added to people's collection when they log in within that window of time. In its announcement, Blizzard said that it will deploy more stability updates, starting with another patch scheduled for release this week. It also said that it's monitoring the game closely for any more issues and bugs that emerge. 

The developer keeps a public list of known issues on its forum, but players are finding more that it has yet to acknowledge. Users are reporting problems regarding specific characters in the game, such as Mei, whose ice wall has been behaving inconsistently, according to Kotaku. Blizzard even had to pull two heroes out of the game completely to address a few bugs in the ability kits. 

Apple's latest Watch Series 8 just came out last month, but you can already grab a deal on it. The 41mm model is selling at Amazon in Midnight or Red for $349, for a savings of $50 (13 percent), and the 44mm model is on sale at $379, or $50 off the regular price. That'll get you all of Apple's latest Watch features, including a skin temperature sensor, low power mode and more. Be aware, though, that stock appears to be limited so you may have to move quickly.

The Series 8 is more of an incremental update over the Series 7, but it does have some useful new features. The first is a new temperature sensor that Apple has tied to women's health, giving female users an estimate on when they may be ovulating. It's meant to be used overnight, sampling your wrist temperature every five seconds so you can see shifts from your baseline temperature. 

The other key feature is Crash Detection. Much as current watches can detect a fall, the Series 8 can detect car crashes via a pair of new accelerometers. It works in concert with the other sensors already included in the Apple Watch to detect four different types of crashes, including rollovers, front impact, back impact and side impact.

While battery life is the same as before at 18 hours, there's a new power mode that keeps it going for up to 36 hours on a full charge. It also uses a newer S8 system-in-package processor, that should allow for improved performance. With those updates, we found the Series 8 to be the "new best smartwatch," letting Apple keep its, er, crown in that department. As mentioned, if you're looking to get one, act quickly.

Follow @EngadgetDeals on Twitter and subscribe to the Engadget Deals newsletter for the latest tech deals and buying advice.