Meta has cut several deals with news publishers to help provide real-time data for its AI chatbot services, as reported by Axios. The commercial agreements will allow its Meta AI chatbots to better answer user queries about news and current events.
These are multiyear deals where publishers will be compensated for the use of their content, but we don't have any monetary specifics. The contracts do stipulate that Meta's chatbots will link out to articles when answering news queries, potentially offering a slight traffic boost to publishers.
The news partners include USA Today, People, Le Monde and CNN. However, there are also a whole lot of conservative outlets included in today's announcement, such as Fox News, The Daily Caller and Washington Examiner. It's a good thing Meta's AI will provide the aforementioned links, just in case a chatbot says something crazy about whatever nonsense culture war topic is going on that day.
This article originally appeared on Engadget at https://www.engadget.com/big-tech/meta-cuts-deals-with-several-news-publishers-for-ai-use-163404107.html?src=rss
The European Commission has fined Elon Musk’s X €120 million (around $140 million) for breaching its transparency rules under the Digital Services Act. The European Union’s executive arm announced that it was investigating the social media company’s blue checkmarking verification system — first introduced when it was still known as Twitter — last year, along with other alleged DSA violations. Today’s verdict concerns the "deceptive design" of the checkmark, as well as "the lack of transparency of [X's] advertising repository, and the failure to provide access to public data for researchers."
The Commission's issue with X’s verification system is that where blue checkmarks were once something that Twitter that Twitter vetted, they can now be bough by anyone. According to the EU, this puts users at risk of scams and impersonation fraud, as they can’t tell if the accounts they’re engaging with are authentic. "While the DSA does not mandate user verification, it clearly prohibits online platforms from falsely claiming that users have been verified, when no such verification took place," it wrote in a statement.
The EU has also ruled that X’s advertisement repository employs "design features and access barriers" that make it difficult for good faith actors and the general public to determine the source of online ads and spot scams or threat campaigns. It says that X fails to provide information pertaining to both the content of an ad and the entity paying for its placement.
The third alleged infringement concerns the public data that companies are required by the DSA to make available to qualifying researchers. The European Commission claims that X’s practices in this area are unnecessarily prohibitive, therefore "effectively undermining research into several systemic risks in the European Union."
X has 60 working days to respond to the EU’s non-compliance decision — the first of its nature — on blue checkmarks, and 90 days to submit an "action plan" of how it will address the alleged breaches relating to its advertising repository and access to public data. Failure to comply could result in financial penalties.
This article originally appeared on Engadget at https://www.engadget.com/big-tech/x-hit-with-140-million-fine-from-the-eu-161259324.html?src=rss
The New York Times and the Chicago Tribune have filed separate lawsuits against Perplexity over alleged copyright infringement. The Times said it had sent Perplexity several cease-and-desist demands to stop using its content until the two reached an agreement, but the AI company persisted in doing so.
In the lawsuit [PDF], the Times accused Perplexity of infringing on its copyrights at two main stages. First, by scraping its website (including in real time) to train AI models and feed content into the likes of the Claude chatbot and Comet browser. Second, in the output of Perplexity's products, with the Times accusing the company’s generative AI products of often reproducing its articles verbatim. The Times also says Perplexity damaged its brand by falsely attributing completely fabricated information (aka hallucinations) to the newspaper.
The Chicago Tribune also filed a lawsuit against Perplexity for similar reasons. "Perplexity’s genAI products generate outputs that are identical or substantially similar to the Chicago Tribune’s content,” the newspaper claimed in its suit. “Upon information and belief, Perplexity has unlawfully copied millions of copyrighted Chicago Tribune stories, videos, images and other works to power its products and tools."
These lawsuits are the latest in dozens of legal cases involving copyright holders and AI companies in the US. The Times, for instance, previously sued OpenAI and Microsoft. It accused the companies of training their large language models on millions of its articles without permission. That case is ongoing.
Copyright holders have licensed their content to AI companies in some cases, though. OpenAI has struck multiple deals with media companies. The Times and Amazon reached an agreement this year that's said to be worth as much as $25 million per year to the media company.
This article originally appeared on Engadget at https://www.engadget.com/ai/the-new-york-times-and-chicago-tribune-sue-perplexity-over-alleged-copyright-infringement-153656431.html?src=rss
Another CES is nearly upon us, another year where we’ll see new gadgets aplenty from giant companies and tiny ones you’ve never heard of. And the not-so-secret secret of CES is that many of these things never make it to market — but usually it isn’t things companies like Samsung show off. But here we are, nearly six years since Samsung first showed off its Ballie personal robot and it is nowhere to be found.
For those who may not recall, Ballie is an adorable circular robot that can putter around your house and project things onto the floor and wall. It’s kind of a virtual assistant on the go. Samsung first revealed this tiny robot at CES 2020, but it was more of a prototype than something anyone expected to purchase. And then there was a global pandemic and we all sort of forgot about weird ball-shaped robots for a few. But Samsung triumphantly unveiled a larger and more refined Ballie at CES 2024, saying it would be on sale that year!
Well, that didn’t happen, but a year later Ballie was back at CES again. Samsung promised it would go on sale in 2025, and followed up with a press release this past April saying it was on track for a summer launch in Korea and the US. As far as I can tell, that’s the last we’ve heard of it.
But with CES looming again, I can’t help but feel like Samsung will roll Ballie out once more, trying to sell the dream of a cute robotic companion who just gets you. I spent some time watching Ballie do its thing in a carefully controlled demo at CES 2024, and I can’t say I was overwhelmed by its purported usefulness or thought there’d be much of a market for this thing. I now can’t help but wonder if Samsung has data backs up my intuition. If this thing was going to sell like gangbusters, it likely wouldn’t be subjected to such a long and public gestation period.
It reminds me a little of one of my favorite Samsung gaffes, the Galaxy Home smart speaker. It was announced at a time when Apple and Google were challenging Sonos and Amazon with voice-activated speakers of their own, moving Siri or the Google Assistant from your phone to a more omnipresent place in your home.
The first rumor of the Galaxy Home happened way back in 2017, and the speaker was officially revealed and briefly shown off by Samsung in August of 2018. My immediate reaction was that this product made very little sense for both Samsung and potential customers — Bibxy sucked, and there were plenty of speakers with better voice assistants. Apparently, Samsung agreed. After multiple years of vague commitments and references to the Galaxy Home, Samsung just… stopped talking about it. Oddly enough, a Galaxy Home Mini speaker was briefly released in South Korea, part of a promotion for people who pre-ordered the Galaxy S20. But I don’t think you could ever just walk into a store and buy one, and the larger Galaxy Home never materialized at all.
Ballie isn’t quite the abandonware situation that the Galaxy Home was, at least not yet. After all, it’s only been about eight months since Samsung dropped that press release claiming it would arrive soon. The company has definitely pushed Ballie in a more public way than the Galaxy Home, making it a little harder to just drop entirely. Maybe we’ll see a revamped Ballie with even more weird tricks next month, or maybe we’ll just get another vague promise that it’ll arrive in 2026. After failing to deliver two years in a row, though, I’m not going to expect Ballie to show up as a real product until I can punch in my credit card and pre-order it... not that I’d do that anyway. Ballie needs to show that it’s a lot more than a cute rolling robot before Samsung gets my cash.
This article originally appeared on Engadget at https://www.engadget.com/ai/where-the-hell-is-samsungs-ballie-robot-151112829.html?src=rss
RAM prices have gone wild, mostly thanks to AI. In this episode, Devindra chats with Will Smith (Brad and Will Made a Tech Pod) about the state of the RAM industry, as well as other hardware we expect to get more expensive. (SSD prices are definitely creeping up too!). Also, we discuss Meta poaching Alan Dye, one of Apple's design executives, and what this could mean for Meta's upcoming devices. And yes, whatever they have next will likely revolve around AI.
Mark Zuckerberg, CEO of Meta, plans deep cuts to his company’s metaverse development – 1:09
Longtime Apple UI designer Alan Dye to join Meta’s AI division – 7:08
US DOT cuts fuel efficiency standards, doubles down on gas cars – 25:40
Waymo autonomous cars recently started driving more aggressively – 31:30
Amazon halts its anime dub beta because it sounded terrible – 38:00
WTF, RAM?? Will Smith joins to talk about why RAM prices are spiraling upward – 44:05
Around Engadget: Metroid Prime 4 is a return to form after 18 years on ice – 1:04:42
Working on – 1:07:36
Pop culture picks – 1:08:32
Credits
Host: Devindra Hardawar Guest: Will Smith Producer: Ben Ellman Music: Dale North and Terrence O'Brien
This article originally appeared on Engadget at https://www.engadget.com/social-media/engadget-podcast-wtf-is-up-with-ram-with-will-smith-from-the-tech-pod-141442002.html?src=rss
Hackers with links to China reportedly successfully infiltrated a number of unnamed government and tech entities using advanced malware. As reported by Reuters, cybersecurity agencies from the US and Canada confirmed the attack, which used a backdoor known as “Brickstorm” to target organizations using the VMware vSphere cloud computing platform.
As detailed in a report published by the Canadian Centre for Cyber Security on December 4, PRC state-sponsored hackers maintained "long-term persistent access" to an unnamed victim’s internal network. After compromising the affected platform, the cybercriminals were able to steal credentials, manipulate sensitive files and create "rogue, hidden VMs" (virtual machines), effectively seizing control unnoticed. The attack could have begun as far back as April 2024 and lasted until at least September of this year.
The malware analysis report published by the Canadian Cyber Centre, with assistance from The Cybersecurity and Infrastructure Security Agency (CISA) and the National Security Agency (NSA), cites eight different Brickstorm malware samples. It is not clear exactly how many organizations in total were either targeted or successfully penetrated.
In an email to Reuters, a spokesperson for VMware vSphere owner Broadcom said it was aware of the alleged hack, and encouraged its customers to download up-to-date security patches whenever possible. In September, the Google Threat Intelligence Group published its own report on Brickstorm, in which it urged organizations to "reevaluate their threat model for appliances and conduct hunt exercises" against specified threat actors.
This article originally appeared on Engadget at https://www.engadget.com/big-tech/chinese-hackers-reportedly-targeting-government-entities-using-brickstorm-malware-133501894.html?src=rss
Portal shooter Splitgate 2 is coming back with a new name and rebuilt experience after being pulled back to beta earlier this year, developer 1047 Games announced. The free-to-play Splitgate: Arena Reloaded will relaunch on December 17th across platforms including Xbox Series X/S, PlayStation 5 and PC.
"After months of community feedback, testing and rebuilding the entire experience, this is Splitgate refocused on what made it special in the first place: classic arena combat without factions, abilities, or extra noise," the developer shared in a blog post, along with a gameplay video.
Splitgate 2 originally launched in June, but 1047 Games pulled it back to beta after deciding that the release had been rushed. For the new version, the company said, "we took the best of Splitgate 1 and Splitgate 2" and cut the parts that didn't fit, while refining the ones that did. "in the process, we not only rebuilt the game, we reconnected with the magic of the genre that built this studio in the first place: a return to the Arena."
Here are some key changes made:
Removed Splitgate 2's factions, abilities, and select equipment so the focus is back on classic arena combat without the extra noise.
Rebuilt progression from the ground up so your time in game feels rewarding, with weapon and character cosmetics you can earn simply by playing.
We've added a true Classic Arena mode with even starts and new map pickups, recapturing the feel of the original.
Fine tuned combat to better match expectations for a modern arena shooter, with more meaningful gunfights and fewer "what just happened" moments.
Completely overhauled the ranked system so your rank reflects your actual skill for players who love to climb and compete.
Added key systems you have been asking for: Mode Select, Player Stats, Ranked Leaderboards, and Account Levels.
Introduced five brand new maps and six fully reworked arenas, bringing the total map pool to 20.
Added three new LMG primary weapons and the iconic Power Weapon, the Railgun.
1047 Games was savaged earlier this year by Splitgate players after co-founder Ian Proulx wore a hat at Summer Games Fest stating "Make FPS Great Again." Later, he posted to the studio's X account on Splitgate 2's release date that he was "not here to apologize" and that the hat was "not a political statement," finally admitting that it was all basically a publicity stunt.
This article originally appeared on Engadget at https://www.engadget.com/gaming/splitgate-arena-reloaded-tries-to-fix-what-splitgate-2-broke-130027714.html?src=rss
Microsoft's Copilot+ initiative launched last year with a clear goal: To produce capable laptops for people eagerly anticipating AI-powered features. Read that sentence again, and it's glaringly obvious that Microsoft's plan was flawed from the start. Most consumers aren't nearly as hyped for AI features as the companies eager to foist artificial intelligence upon us. And those features aren't exactly compelling, either. Microsoft's Recall — which snaps screenshots of your PC to create a database of everything you’ve done– was dogged by privacy concerns from the start. And to be honest, I haven't found its ability to remember the files and websites I've opened to be that useful.
Without any sort of killer AI app, most consumers weren't going to pay a premium for Copilot+ systems either. Not in this precarious economy, anyway. So it wasn't a huge surprise to see sales of Copilot+ systems going practically nowhere over the last year. In the third quarter of 2024, they accounted for less than 10 percent of systems shipped, according to data from Mercury Research (via Tom’s Hardware). The research firm IDC (via PCWorld) also found that Copilot+ systems made up just 2.3 percent of Windows machines sold in the first quarter of 2025 (and a mere 1.9 percent of the entire PC market).
Instead of continuing to promote Copilot+, Microsoft now wants to "make every Windows 11 computer an AI PC". The new "Hey Copilot" voice commands and Copilot Vision, a feature that lets the AI assistant see what's on your screen, are both cloud-powered. That means you won't need the beefy 40 TOPS neural processing units (NPU) found on Copilot+ systems to use them. Microsoft spent the past few years touting NPUs as the gateway to useful AI features, like Recall and Windows Studio webcam effects, but only one of its new AI capabilities actually requires an NPU. (And even that is just a slight update to Click to Do, allowing you to send Zoom invitations by right-clicking on e-mail addresses.)
It's easy to view the whole Copilot+ initiative as a cynical way to ramp up AI hype and push people towards expensive new laptops, especially as the October 14 Windows 10 end of support date loomed. But it also led to some genuinely useful changes: Microsoft made 16GB of RAM a standard for Copilot+ systems, along with 256GB of storage and the aforementioned 40 TOPS NPUs. The launch of Copilot was also the kick in the pants Microsoft needed to revamp Windows for mobile Arm processors. I never thought I'd love a Surface with a Snapdragon chip, but the improved Arm support on the Surface Pro and this year's smaller model finally won me over.
The Dell 16 Premium sitting on a ledge.
I wouldn’t call the Copilot+ program a huge swing, but it’s still the sort of industry-wide cat herding that’s rare to see in the PC space. Microsoft couldn’t just snap its fingers and shift all PCs to efficient mobile chips with powerful NPUs, like Apple did with its own jump to M-series chips years ago. Microsoft had to wait for new NPU-equipped hardware from Qualcomm (and eventually Intel and AMD). It had to finally fix the Windows on Arm problem. And it also had to double-down on AI features that felt truly transformative. It’s just a shame that consumers didn’t seem to care.
Microsoft said that Copilot+ systems accounted for 15 percent of premium PCs sold during last year’s holiday season, but the company hasn’t released any new sales figures since then. “This is the fastest adoption I've seen of a new category of hardware, and we've done it faster than the normal generational shift of silicon,” James Howell, Microsoft’s VP of Windows marketing, said in a conversation with Engadget. “Copilot+ PCs continue to be a transition that we are pushing for and prioritizing. But I can't give you the exact numbers beyond that… Just for the last two or three months, we've been doing pretty well with year-on-year growth in the Windows business.”
Surface Pro Copilot+
Devindra Hardawar for Engadget
While Microsoft ultimately doesn’t have much to show for the Copilot+ initiative, the steady progression of hardware will lead to AI PCs dominating over the next five years. The research firm Omdia predicts that AI PCs will account for 55 percent of computers shipped in all of 2026, up from 42.5 percent of systems in Q3 2025. By 2029, Omdia predicts AI PCs will make up 75 percent of all systems shipped, giving Windows 80 percent of the AI PC market.
Omdia AI PC shipment predictions
Omdia
“It’s important to note that this steep adoption curve [for AI PCs] is driven more by the product roadmaps of the PC market, rather than consumers and businesses seeking PCs specifically for AI,” according to Omdia research analyst Kieren Jessop. “For businesses, and consumers especially, AI-capable PC adoption is more a function of a customer going to purchase a device and that device just so happens to have an NPU.”
Microsoft was basically right: AI PCs are the future. But it turns out the AI features people actually want to use — like ChatGPT, Sora and Microsoft’s own Copilot — are mostly powered by the cloud, making onboard NPUs superfluous. That won’t be true forever. There are tangible security, speed and convenience benefits for onboard AI processing, like transcribing sensitive audio instead of sending it to the cloud. But for now, those AI workloads are relatively niche, and they’re not enough to make the Copilot+ a true success by any measure.
This article originally appeared on Engadget at https://www.engadget.com/computing/laptops/microsofts-copilot-ai-pc-plan-fizzled-but-it-still-served-a-purpose-130000239.html?src=rss
Spinning off from the action-camera company Insta360, Antigravity now has its debut drone on sale. With 360-degree cameras that capture 8K and offer you a truly unconstrained view of the skies, the A1 is a different drone from everything else out there. Sorry, DJI.
Instead of typical drone joysticks, you get a motion controller that lets you point and shoot like video game gesture controls, while crisp FPV goggles put you right inside the cockpit.
It’s easy to fly after takeoff, but the A1’s myriad parts are often tricky to sync together — and pulling video down to the companion app is even trickier. Going on specs alone, like speed and camera sensor size, it doesn’t stand up to cinematic drones from the likes of DJI.
Still, it’s not meant to be a cinematic drone. It’s a hybrid mix of flight experience, FPV drone and a not-miss-a-thing camera drone. It’s truly unique — and fun.
Amazon has quietly removed its terrible AI-generated English dubs for several anime shows on Prime Video, following widespread ridicule from viewers and the industry. AI dubs were recently added to Banana Fish, No Game, No Life and Vinland Saga, where they were labeled “AI beta” in the Languages section of the app.
For shows lacking an English-language dub, it was a seemingly cheap way to consume anime for Amazon. However, it quickly became clear that the dubs were really quite bad. Baaaad.
Voice actor Daman Mills called the AI-generated dub for Banana Fish a “massive insult to us as performers” in a post on X.
The company continues to invest heavily in its own shipping network.
An Amazon double today. According to The Washington Post, Amazon is considering discontinuing use of the US Postal Service and building its own shipping network to rival it. The e-commerce behemoth spends more than $6 billion a year on the public mail carrier — almost 8 percent of the service’s total revenue. That’s up from just under $4 billion in 2019. That split might be due to a breakdown in negotiations between Amazon and the USPS rather than Amazon proactively pulling its business.
Amazon has invested heavily in all kinds of delivery methods, including shipping logistics, buying its own Boeing planes, launching its own electric delivery vans and slowly building a drone delivery network.
A triple? Sorry. Amazon didn’t have a specific release date to share beyond “later this year” for its latest Scribe slates. And talk about brinkmanship! Here we are in December. The company says the devices will be available on December 10. This is the third generation of the Kindle Scribe line of E Ink writing tablets — the first time Amazon has three versions of the Scribe. At the entry level, the Scribe without a front light starts at $430, while the model with a light starts at $480. The Kindle Scribe Colorsoft will start at $630. You always have to pay more for color.
A highly capable cinema camera at a reasonable price.
The Nikon ZR could be a breakthrough for content creators, largely because it incorporates technology from RED — a company now owned by Nikon. The combination of professional-grade video quality (specifically RED RAW) and autofocus comes at a fraction of the cost of dedicated cinema rigs. There are some compromises on battery life and the lack of a viewfinder, but the ZR arguably offers the best video quality for the money.
In a statement, Netflix said it expects to "maintain" Warner Bros. current operations, as well as its policy of theatrical releases for its films. But the deal may spell the end for HBO Max as its own product in the longer term, as the statement also says "by adding the deep film and TV libraries and HBO and HBO Max programming, Netflix members will have even more high-quality titles from which to choose."
Naturally, the deal will see Netflix become one of the biggest players in global media, combining its global reach with some of the most recognizable names in entertainment. That includes HBO, DC Studios, Cartoon Network, its game development studios and TCM, as well as the chunks of TNT not cast adrift with Discovery.
It's likely the deal will not go ahead without a lot of objections from other buyers, as well as the government itself. Yesterday, Paramount Skydance said (via the Hollywood Reporter) any deal between WB and Netflix would be the result of an "unfair" process. Given the close ties between Paramount's new owners and the administration, it's likely any deal will be subject to scrutiny as well as the usual questions around the size of the combined operation.
Since the announcement was made, Engadget senior reporter Devindra Hardawar has spoken with Hollywood players and collated studies and statements to answer any burning questions you might have on what this deal means for you. He also answers questions about the likelihood of regulatory approval, theatrical releases and physical media. Catch up on all that in his piece titled “The Netflix and Warner Bros. deal might be great for shareholders, but not for anyone else.”
Update, December 5 2025, 1:45PM ET: This story has been updated to add a paragraph and link to a new article we’ve published that contains deeper analysis and more information about the Netflix/Warner Bros. deal and what that might mean for streaming, movies, TV and shareholders.
This article originally appeared on Engadget at https://www.engadget.com/entertainment/netflix-to-buy-warner-bros-for-827-billion-120836295.html?src=rss
