Max is putting you in the driver’s seat for NASCAR’s upcoming races. The streamer will now let you watch up to four in-car driver cams at once, starting with the Cook Out Clash on Sunday, February 2nd.
You can switch between up to 40 different driver cams, with each 1080p feed offering the sounds of both the scanner team radios and ambient car noise. If you’re having trouble choosing which drivers to watch in multiview, Max will offer preset options “based on the most compelling matchups.”
Though Max won’t present a top-down view of the race, it will still show a leaderboard, along with the lap number, race status, and stage while you watch the driver cams. You can also switch between the audio on each feed while you’re watching in multiview.
The in-car cams will be available exclusively on Max throughout this year’s NASCAR season. You can access the feeds across all of Max’s platforms, whether you’re watching on a TV or your phone.
“This is such a great way for fans to complement the television experience or the broadcast experience,” Hania Poole, Warner Bros. Discovery Sports’ SVP of digital and product, said in an interview with The Verge. “I think that this product specifically will just continue to evolve, and we’ll iterate, we’ll learn, and we’ll test, and maybe apply it to other scenarios.”
The Chinese startup DeepSeek shook up the world of AI last week after showing its supercheap R1 model could compete directly with OpenAIâs o1. While it wiped nearly $600 billion off Nvidiaâs market value, Microsoft engineers were quietly working at pace to embrace the partially open- source R1 model and get it ready for Azure customers. It was a decision that came from the very top of Microsoft.
Sources familiar with Microsoftâs DeepSeek R1 deployment tell me that the companyâs senior leadership team and CEO Satya Nadella moved with haste to get engineers to test and deploy R1 on Azure AI Foundry and GitHub over the past 10 days. For a corporation the size of Microsoft, it was an unusually quick turnaround, but there are plenty of signs that Nadella was ready and waiting for this exact moment.
While the open-source model has upended Wall Streetâs idea of how much AI costs, Nadella seemed to know that something like DeepSeek was coming eventually. Appearing on the BG2 podcast in early December, he warned of the exact thing DeepSeek went on to achieve weeks later: an algorithmic breakthrough that results in compute efficiency.
How this breakthrough has been achieved is …
Generative artificial intelligence output based purely on text prompts — even detailed ones — isn’t protected by current copyright law, according to the US Copyright Office.
The department issued this guidance in a broad report on policy issues regarding AI, focused on the copyrightability of various AI outputs. The document concludes that while generative AI may be a new technology, existing copyright principles can apply without changes to the law — and these principles offer limited protection for many kinds of work.
The new guidelines say that AI prompts currently don’t offer enough control to “make users of an AI system the authors of the output.” (AI systems themselves can’t hold copyrights.) That stands true whether the prompt is extremely simple or involves long strings of text and multiple iterations. “No matter how many times a prompt is revised and resubmitted, the final output reflects the user’s acceptance of the AI system’s interpretation, rather than authorship of the expression it contains,” the report says.
This decision would seemingly rule out protections for works like “Théâtre D’opéra Spatial,” a controversial award-winning Midjourney-generated image whose creator fought an extended battle to register it with the Copyright Office.
The office demonstrates the unpredictability of AI systems with a Gemini-produced image of a cat smoking a pipe and reading a newspaper, noting that Gemini ignored some prompt instructions and added a few things of its own — including an “incongruous human hand.” It contrasted this process with Jackson Pollock’s splatter painting method, where he didn’t control the exact placement of the paint on the canvas, but “controlled the choice of colors, number of layers, depth of texture, placement of each addition to the overall composition — and used his own body movements to execute each of these choices.” Ultimately, the office writes, “the issue is the degree of human control, rather than the predictability of the outcome.”
“No matter how many times a prompt is revised and resubmitted, the final output reflects the user’s acceptance of the AI system’s interpretation, rather than authorship of the expression it contains.”
At the same time, the Copyright Office says that simply using AI to assist in human creative output does not necessarily jeopardize that work’s ability to be protected by the law. There’s a difference between AI being used as a tool to assist a creative work and “AI as a stand-in for human creativity,” and the office says that further analysis is warranted. But it assures creatives that using AI to outline a book or come up with song ideas shouldn’t impact the ability to copyright the final human-produced work, since the author is simply “referencing, but not incorporating, the output.”
Artists can get some protection if they feed their own work into an AI system for modification — by, say, using a tool to add 3D effects to an illustration. AI-generated elements of the work still wouldn’t be protectable, but if the original product remains recognizable, the “perceptible human expression” in the work could still be covered by copyright.
People can also receive protection for works that incorporate AI-generated content as long as there’s significant creative modification. A comic with AI-generated images can be covered if a human arranges those images and pairs them with human-written text, though the individual AI-generated images wouldn’t be protected. Likewise, “a film that includes AI-generated special effects or background artwork is copyrightable, even if the AI effects and artwork separately are not.” On a “case-by-case determination,” even prompt-generated images could be protected if a human selects and remixes specific areas of the picture. The office compares these scenarios to making copyrightable derivative works of human-created art — minus the original human.
A separate question is whether text prompts themselves can be protected by copyright. Overall, the office compared prompts to “instructions” that convey uncopyrightable ideas, but it acknowledged some particularly creative ones could include “expressive elements.” This doesn’t, however, translate into the work they produce being protected.
The Copyright Office didn’t rule out the potential for this to change if the technology evolves. “In theory, AI systems could someday allow users to exert so much control over how their expression is reflected in an output that the system’s contribution would become rote or mechanical,” the report says. But as of now, it doesn’t seem that prompts “adequately determine the expressive elements produced, or control how the system translates them into an output.”
This document is part of a larger effort by the Copyright Office to clarify policy questions and identify legal gaps around AI, starting with a July 2024 report encouraging new deepfake laws. The office next plans to issue a third and final report on its findings on “the legal implications of training AI models on copyrighted works.”
Microsoft is bringing OpenAI’s o1 reasoning model to all Copilot users this week. You won’t need to subscribe to a $20 monthly Copilot Pro or ChatGPT Plus plan to get it either, as Microsoft is making it free for all users of Copilot.
Think Deeper, as Microsoft calls its integration of o1, works by allowing Copilot to handle more complex questions. You can tap the Think Deeper button inside Copilot, and it will take around 30 seconds to “consider your question from all angles and perspectives.”
Microsoft first launched Think Deeper in October, providing a preview of the feature inside Copilot Labs, which lets Copilot Pro subscribers experiment with new features that Microsoft is developing. Much like ChatGPT Plus, Think Deeper will supply step-by-step answers to complex questions, so it’s good for comparing two options, creating code for apps, or planning a long road trip.
Microsoft AI CEO Mustafa Suleyman revealed that the company will now offer Think Deeper at no extra cost to all Copilot users in a LinkedIn post yesterday. “I’m genuinely so excited that our tens of millions of users are all getting this opportunity,” says Suleyman. “We’ve got so much more in the pipeline right now that I can’t wait to tell you about.”
The Federal Aviation Administration is facing its first major aviation disaster in 16 years without a leader because Elon Musk helped push him out.
Michael Whitaker stepped down as FAA administrator on January 20th, the day of Donald Trump’s inauguration, after clashing with Elon Musk, whose company SpaceX is regulated by the agency. Musk publicly called on Whitaker to resign after the FAA fined SpaceX for failing to get approval for launch changes. And now the power vacuum at the agency is coming into sharp focus after an Army helicopter collided with an American Airlines jet Wednesday evening in Washington, DC, killing everyone on board.
At a press conference Thursday with rescue officials, newly confirmed Transportation Secretary Sean Duffy ignored reporters’ questions about whether the FAA had an acting director in place to help manage the crisis.
REPORTER: Is there an acting FAA director?DUFFY: *walks away*(The FAA director resigned on the day of Trump's inauguration under pressure from Elon Musk)
Whitaker was only a year into the job when he announced his intention to step down. He had several years left in his term. His resignation clears the way for President Donald Trump to name his own replacement to run the agency. Later Thursday morning, Trump said he was tapping Chris Rocheleau, a top executive for an aviation business association, as acting FAA administrator, but provided no details about a permanent replacement.
But Musk’s efforts to get Whitaker were well known even before Trump’s victory in November. He has complained many times about the FAA, lashing out in September after the agency levied a $633,000 fine for launching missions with unapproved changes. (Musk is worth over $400 billion, making him the richest man in the world.)
The FAA has also fined Starlink, after the SpaceX subsidiary failed to submit safety data before launching satellites in 2022. In a House hearing, Whitaker explained that the FAA’s civil penalties were “the only tool we have to get compliance on safety matters.”
On X, Musk complained that the FAA was “harassing SpaceX about nonsense that doesn’t affect safety while giving a free pass to Boeing even after NASA concluded that their spacecraft was not safe enough to bring back the astronauts.” He also claimed that humans would never land on Mars without “radical reform at the FAA.” In September, he wrote “he needs to resign” about Whitaker.
In addition to Musk’s attacks, the FAA has long suffered from underfunding and outdated technology. In 2023, a panel of experts released a report that found that the increasing reliance of overtime to staff air traffic control facilities was putting air safety at risk. The agency has fielded hundreds of complaints from air traffic workers describing dangerous conditions from staff shortages to dilapidated buildings.
Update, January 30th: After this story was publish, Trump announced that he was naming Chris Rocheleau as acting FAA administrator. This story has been updated to reflect that fact.
On today’s episode of Decoder, we’re talking about the only thing the AI industry — and pretty much the entire tech world — has been able to talk about for the last week: that is, of course, DeepSeek, and how the open-source AI model built by a Chinese startup has completely upended the conventional wisdom around chatbots, what they can do, and how much they should cost to develop.
DeepSeek, for those unaware, is a lot like ChatGPT — there’s a website and a mobile app, and you can type into a little text box and have it talk back to you. What makes it special is how it was built. On January 20th, the startup’s most recent major release, a reasoning model called R1, dropped just weeks after the company’s last model V3, both of which began showing some very impressive AI benchmark performance. It quickly became clear that DeepSeek’s models perform at the same level, or in some cases even better, as competing ones from OpenAI, Meta, and Google. Also: they’re totally free to use.
But here’s the real catch: while OpenAI’s GPT-4 reported training cost was as high as $100 million, DeepSeek’s R1 cost less than $6 million to train, at least according to the company’s claims. In a matter of days, DeepSeek went viral, becoming the No. 1 app in the US, and on Monday morning, it punched a hole in the stock market.
Panicked investors wiped more than $1 trillion off of tech stocks in a frenzied selloff earlier this week. Nvidia, in particular, suffered a record stock market decline of nearly $600 billion when it dropped 17 percent on Monday.
For more than two years now, tech executives have been telling us that the path to unlocking the full potential of AI was to throw GPUs at the problem. Since then, scale has been king. And scale was certainly top of mind less than two weeks ago, when Sam Altman went to the White House and announced a new $500 billion data center venture called Stargate that will supposedly supercharge OpenAI’s ability to train and deploy new models.
The aftermath has been a bloodbath, to put it lightly. Venture capitalist Marc Andreessen sounded the alarm, calling DeepSeek “AI’s Sputnik moment” — and that does appear to be how the AI industry and global financial markets are treating it.
In DeepSeek and Stargate, we have a perfect encapsulation of the two competing visions for the future of AI. One is closed and expensive, and it requires placing an ever-increasing amount of money and faith into the hands of OpenAI and its partners. The other is scrappy and open source, but with major questions around the censorship of information, data privacy practices, and whether it’s truly as low-cost as we’re being told.
What is clear is that we’ve entered a new phase in the AI arms race, and DeepSeek and Stargate represent more than just two distinct paths toward superintelligence: they also represent a new, escalating front in the US-China relationship and the geopolitics of AI. This is becoming especially fraught, as President Donald Trump continues to wreak havoc on foreign relations with a new threat of tariffs on foreign semiconductors.
There is a whole lot going on here — and the news cycle is moving very fast. So to break it all down, I invited Verge senior AI reporter Kylie Robison on the show to discuss all the events of the past couple weeks and to figure out where the AI industry is headed next.
If you’d like to read more about what we talked about in this episode, check out the links below:
Why everyone is freaking out about DeepSeek | The Verge
Valentine’s Day might seem like any other commercial holiday given the abundance of candy and the sheer number of cards lining store shelves, but if you put in a little effort and a lot of heart, it can be about so much more. After all, at its core, February 14th is about celebrating love in all its forms, with or without sweet treats.
One way to show your appreciation for your loved one — whether it be your partner, sister, parent, or best friend — is with a gift that speaks to who they are. Luckily, if you need some inspiration, the Verge staff has pooled together a list of thoughtful gift ideas. Some of these, such as the AirPods 4 and latest Kindle Paperwhite, are gadgets we use on the daily, while items like Amazfit’s Active 2 and Soundcore’s latest speaker are more recent finds we’re still putting through their paces.
We’ve also sprinkled in a few ideas for those who could use a little less tech in their lives, so there should be something for everybody on your list. After all, nothing says “I love you” quite like a scratcher.
The rumors are true. Microsoft is launching Intel-powered versions of its Surface Pro 11 and Surface Laptop 7. Available on February 18th, both of these new Surface devices are Copilot Plus PCs, but instead of using Qualcomm’s Arm-based Snapdragon X series of chips, they’ll use Intel’s Lunar Lake processors.
Microsoft is targeting these models at businesses, much like how it released Intel-only versions of the Surface Pro 10 and Surface Laptop 6 last year. Business customers will now be able to pick between Intel or Qualcomm inside Surface Pro 11 and Surface Laptop 7 designs that are identical.
The Surface Laptop 7 for businesses will be available starting at $1,499.99, which is $500 more than the starting price of the Qualcomm variant. Two different screen sizes will be available: 13.8-inch and a larger 15-inch model, and both can be configured with Intel Core Ultra 5 or Ultra 7 chips, up to 32GB of RAM, and up to 1TB of storage. Microsoft is promising up to 14 hours of battery life (based on active web usage) on the 15-inch model and up to 12 hours on the 13.8-inch variant. These estimates are only an hour behind the same figures Microsoft provides for the Qualcomm versions of the Surface Laptop 7.
The ports on the Intel-powered Surface Laptop 7 are largely unchanged, with Microsoft opting for a USB-A 3.2 port instead of the USB-A 3.1 that ships on the Qualcomm version. There are also two USB-C 4 / Thunderbolt 4 ports, the Surface Connect charging port, and a headphone jack. Microsoft is also using the same PixelSense displays on the Intel variant of the Surface Laptop 7 as the Qualcomm version.
Much like the Surface Laptop 6 for businesses, Microsoft is shipping an optional smart card reader in certain markets. Microsoft is also launching a 5G version of the Surface Laptop 7 later this year, and the company plans to share more details on this model “in the coming months.”
The Surface Pro 11 for businesses is also largely identical to the Qualcomm version, but it starts at $1,499.99 instead of the $999.99 starting price of the Qualcomm version. You can pick between a 13-inch LCD or OLED panel, with both displays the same as on the Qualcomm-powered Surface Pro 11. Businesses will be able to configure the Intel-powered Surface Pro 11 with Core Ultra 5 or Ultra 7 processors, up to 32GB of RAM, and up to 1TB of storage.
Microsoft hasn’t made any changes to the ports on the Intel Surface Pro 11, so there are two USB-C 4 / Thunderbolt 4 ports and a Surface Connect for charging. If you want to use headphones on the Surface Pro 11, you’ll have to use Bluetooth-powered ones or connect a wired pair over USB-C. The Surface Pro 11 also has a built-in NFC reader to support security keys like the YubiKey 5C NFC.
Both of these Intel-powered Surface devices come just months before Windows 10’s end of support on October 14th. Microsoft is now pitching these Copilot Plus PCs as upgrade devices for businesses that are transitioning from Windows 10 to Windows 11. As they both include NPUs capable of Copilot Plus features, they’ll be able to access the latest Windows AI features including Recall and Microsoft’s AI-powered Windows Search.
Microsoft is also tempting businesses with a new USB 4 dock that will also be available on February 18th, priced at $199.99. Designed to work with all Surface devices with a USB-C port, the dock supports two 4K displays at 60Hz and includes 65-watt passthrough charging and up to 40Gb/s for data transfers. Front-facing ports include a single USB-A 3.2 Gen 2 port and a USB-C 4 Gen 3 port, while at the rear, there is a single USB-C 4 Gen 3 port, an HDMI 2.1 port, an ethernet port, and a USB-C connector for the 100-watt power supply.
Microsoft is opening up business preorders for the Intel-powered Surface Pro 11 and Surface Laptop 7 today, and both devices ship on February 18th.
At Appleâs September 2024 hardware event, the company announced its iPhone 16 lineup, which was âbuilt for Apple intelligenceâ â but Appleâs leap into AI has had a soft landing. The phones arrived without Apple Intelligence and, over the course of four months, Apple has been slowly rolling out some of the promised features through iterations of iOS 18. Â
Apple Intelligenceâs first set of features included tools weâre pretty familiar with, like glowing gradients to indicate the presence of AI, the ability to change the tone of an email, and a photo clean-up tool that lets AI eliminate distractions. Now, weâre up to iOS 18.3 with updates that include Visual Intelligence, Image Playground, AI image generation, Genmoji, and a ChatGPT extension for Siri. Â
This isnât the end of AI for iOS; there are even more updates coming with iOS 18.4 that might allow Siri to understand whatâs on your screen and take action in apps for you. But right now, is Apple Intelligence even close to delivering on what was announced in September? Check out our recap to see where we are now.
The smart home holds so much promise. It can make life more convenient with lights that turn on as you walk in a room, doors that unlock as you approach, and robots that clean your floors. It can also make your home safer, more energy efficient, and even a little more fun. (Have you tried asking Alexa to beam you up?)
But for all its benefits, the smart home can be complicated, confusing, and occasionally maddening. It’s also hard to keep up with all the changes. New gadgets are arriving daily, new features come to old products, and there are so many different ways to turn on a smart light bulb.
If you need a guide, that’s what I’m here for.
Here, I’ll be posting the latest smart home reviews, guides, news, and opinions on everything happening in the connected home. Follow this page to stay updated on what Apple, Amazon, Samsung, Google, and Home Assistant and the rest are doing with their smart home platforms. I’ll keep you in the loop on all the newest technologies — including Matter, Thread, Sidewalk, UWB — as well as the old favorites. And, of course, I’ll cover all the news on the latest gadgets and the biggest releases around tech for your home.
One month into the year and Netflix is ready to detail its plans for 2025 — or at least, some of them. In a now-annualtradition, the streamer has outlined some of its biggest releases across film, television, and gaming.
And while plenty of the listed titles were already known, there are also some pleasant surprises. The movie lineup looks particularly strong, with Wake Up Dead Man — the next Knives Out mystery, following Glass Onion — and Guillermo del Toro’s take on Frankenstein both due out later this year. Meanwhile, some of the streamer’s most important properties are coming to an end with the final seasons of both Stranger Things (which still doesn’t have a premiere date) and Squid Game (which now does).
And while Stranger Things is ending, its creators, the Duffer brothers, did detail two projects they will have coming to Netflix in 2026: The Boroughs and Something Very Bad Is Going to Happen. (They also hinted that there are some Stranger Things spinoffs in the works, without getting into detail.)
As for what to expect this year, here are the highlights:
Film
The Witcher: Sirens of the Deep — February 11th
Plankton: The Movie — March 7th
The Electric State — March 14th
The Old Guard 2 — July 2nd
Fear Street: Prom Queen — summer
Frankenstein — November
Wake Up Dead Man: A Knives Out Mystery — fall
RIP — fall
Happy Gilmore 2 — 2025
K-Pop: Demon Hunters — 2025
TV
Cobra Kai season 6 (part 3) — February 13th
Zero Day — February 20th
Devil May Cry — April 3rd
You season 5 — April 24th
Squid Game season 3 — June 27th
Black Mirror season 7 — 2025
Department Q — 2025
The Sandman season 2 — 2025
Stranger Things season 5 — 2025
The Witcher season 4 — 2025
Wayward — 2025
Wednesday season 2 — 2025
Games
While the lineup of (announced) games coming to Netflix is fairly small, there are two notable highlights: a mobile version of WWE 2K this fall to go along with the new deal between the two, and more Netflix Stories games (Netflix’s branding for interactive fiction) based on series like Ginny & Georgia, Sweet Magnolias, Love Is Blind, and Outer Banks.
It was a long three-year wait between the first seasons of Squid Game, but luckily, the finale is coming much faster. Netflix announced today that season 3 of Squid Game — which will also be the show’s last — will start streaming on June 27th.
That’s around six months after the long-awaited season 2 premiered on Netflix on December 26th of last year. The show, which has gone on to become Netflix’s biggest series, originally kicked off in 2021. There aren’t a lot of details for what viewers can expect in season 3, but Netflix did release a handful of images as part of the announcement, which show some of the survivors from last season’s bullet-soaked finale:
Nvidia is launching its next-gen RTX 5090 and RTX 5080 GPUs today, but it’s also releasing an exciting update for existing RTX GPU owners. A new GPU driver (572.16) allows you to force DLSS 4 inside games or apps that don’t currently support it, providing improved image quality and even less VRAM usage in some cases.
While DLSS 4 is debuting today in 75 games and apps, the Nvidia app now has a new override feature that will improve image quality in games that use DLSS by allowing you to switch to Nvidia’s new transformer model. This means you don’t have to wait on developers to update their games. Owners of RTX 40-series GPUs will also get access to a new frame generation model that is 40 percent faster and uses 30 percent less VRAM than the old one.
Nvidia has also updated its RTX Video Super Resolution (VSR) feature, which can upscale old blurry YouTube videos, to a new and more efficient model. It now uses 30 percent fewer GPU resources on the highest-quality setting, allowing more RTX GPUs to enable this feature. VSR will also now upscale HDR video, so if you’re watching lower-resolution HDR videos, these can now be automatically upscaled.
While existing RTX owners will get access to more efficient models, those looking to purchase RTX 50-series cards today will also be able to enable DLSS Multi Frame Generation. This new feature builds on top of Frame Generation by generating three additional frames per traditionally rendered frame.
In our RTX 5090 and RTX 5080 reviews, we found that Multi Frame Generation is a promising addition but only works well if your base frame rate after DLSS Super Resolution is applied is high. Otherwise, even if your base 30fps after Super Resolution gets boosted beyond 100fps, it will still have the input latency of a 30fps game.
Nvidia also surprised potential RTX 50-series owners with a new Smooth Motion feature in today’s driver update. It’s a driver-based AI model that enables “smoother gameplay by inferring an additional frame between two rendered frames” for games without DLSS Frame Generation support. Nvidia says Smooth Motion is exclusive to the RTX 50-series and can be applied to games running with DLSS Super Resolution, at native resolution, or even titles with other upscaling technologies and it will typically double “the perceived frame rate.”
We’ve got several parents on our staff, and as members of The Verge, they are also into gadgets. Put the two together, and you’ve got some really interesting stuff that can be used to keep your infant, toddler, or child safe, happy, educated, and / or out of trouble.
Here are some of the devices that the folks here use to help be better — and happier — parents.
Infants
The Yogasleep Hushh is a small, portable white noise machine. That’s it, and that’s why it’s great.
It’s not a smart device and doesn’t require any sort of subscription. It has physical buttons. It has three different white noise options. It can run on battery for up to 24 hours, and you can charge it with a USB-C cable. It even comes with a ring to attach it to things.
My wife and I turn on the Hushh every time we put our baby down to sleep. If you’re looking for a simple white noise machine that just works, this is the one to buy. – Jay Peters, news editor
I think Vava has one of the only no-gimmicks baby monitors right now that has the few things I wanted: a big screen, good battery life, zero internet features, and USB-C charging. When my wife and I had our first child just over a year ago, it felt like our short wishlist was impossible to find in the market, regardless of price. It has never failed to connect even at the furthest point of our home and has taken many tumbles without breaking. We tried the Panasonic KX-HN4101W that had a built-in sound machine and charged with Micro USB, but it would not turn on after a few months of use. It did have a very sensitive motion detection feature that was only useful enough to feed into my anxieties as a new parent. – Umar Shakir, news writer
Much like Umar’s pick above, the Infant Optics DXR-8 Pro my wife and I opted for is a no-fuss 720p video monitor requiring no internet connection. One of its advantages is you can buy a cheap wide-angle lens for it that easily gives you a full view of a crib or bassinet. That lens was clutch at family get-togethers, where we had to set up the little one to nap in a spare bedroom with cramped spacing. My only gripes are the use of barrel plugs instead of USB-C and that you can only crop into the center of the frame when you have those moments of illogical worry and want to make sure the baby is breathing.
I think the separation these non-connected monitors offer, compared to cloud-connected ones, can be a healthy one. It may not be for everyone, but if one parent is home with the baby while the other is out grocery shopping or actually socializing, it not only prevents judgy snooping but also helps them be present where they are outside the house. – Antonio G. Di Benedetto, reviewer
CRKD’s Atom is not a baby toy at all, but this wireless mini controller is certainly baby-sized. I don’t want to force my interests onto my kid, but even before her first birthday, the little one has already taken to picking up some of our various game controllers around the house and fidgeting with the buttons and sticks. We’ve since designated a couple of older gamepads for baby plaything duty, prompting me to try putting the tiny Atom in her hands. Along with some Xbox and PlayStation controllers, the Atom is now in rotation for her to carry around and press its satisfyingly clicky D-pad and buttons.
I imagine this little gamepad could be an accessible on-ramp to baby’s first video games, and I even put it to the test one day. I fired up Geometry Wars: Retro Evolved on the TV (via a docked Steam Deck), customized the controls to just use the D-pad, and watched her gently move around and shoot onscreen. She didn’t have much of an idea of what was going on, but she enjoyed the bright colors and identified that she was controlling it before enemies blew her up in seconds. (I need an invincibility mode for her.) Then my self-awareness and “Oh shit, is this bad parenting?” senses kicked in, and I turned off the game. Maybe I’ll just stick to her mimicking dad and mom playing on her unpaired controller for now. – Antonio G. Di Benedetto, reviewer
When you have a kid, you’re not just gaining another human. You’re also gaining all of the hardware needed to care, feed, and even transport them. The GB Pockit Air won’t fit all of your stroller needs since it’s only designed for kids ages six months and up, but for those times when you’re trying to travel light, you’ll appreciate its clever engineering. It weighs just 10.1 pounds and folds down small enough to fit in a backpack or shoulder bag, and it’s airplane hand luggage-compliant. Despite its lightweight design, the Pockit Air still features eight wheels for stability, padded shoulder straps, and a mesh pocket on the bottom for carrying other kid necessities. – Andrew Liszewski, senior reporter, news
The Snoo is a smart bassinet that plays white noise and rocks your baby, ramping up the intensity if your baby keeps crying. I have heard from friends that it does not work for every baby, but when it works, it really works — and you have truly magical moments when it successfully puts your baby to sleep. You can control the intensity manually and get nice sleep logs in their proprietary app — though Snoo’s parent company, Happiest Baby, has come under fire for introducing a new subscription fee of $19.99 / month for the main app functionality after nine months. While infuriating, I was so desperate for sleep, I would still have bought a Snoo. Also look out for return policies; some parents prefer to buy their Snoos from Amazon to avoid Happiest Baby’s steep $199 restocking fee. – Helen Havlak, publisher
These have been my son’s favorites — music, in particular, is a big hit with him. Any musical toy may eventually drive you insane, but these cycle through enough tunes that I am still dancing along to them with my son. The Fisher-Price Flap & Wobble Penguin is a musical penguin toy that cycles through surprisingly boppy remixes. It flaps its wings with high-contrast black-and-white spots that will mesmerize your baby. Meanwhile, the Take Along Tunes Musical Toy is a great cheap little teether and music-maker that has amused my baby for hours. And at less than $10, it is a good deal. – Helen Havlak, publisher
If you plan to pump or to formula feed, it’s nice to have an easy way to warm bottles. Our baby will drink room-temperature bottles, but he definitely prefers nicely warmed milk (who wouldn’t?). This warmer sits on our counter, looks reasonably nice, and works well. – Helen Havlak, publisher
Toddlers and kids
Before my kid’s hands were large enough to hold a Nintendo Switch controller, this kid-sized handheld satisfied their gaming curiosity. What’s unique about the LeapFrog RockIt Twist is that it can be held four different ways, with each side of the handheld featuring unique control mechanisms like dials, a slider, spinners, various buttons, and a more traditional D-pad. It comes with 12 preloaded games that utilize the unique controls while also covertly teaching skills like math, problem-solving, and literacy so you’ll feel slightly less guilty about all the screen time. – Andrew Liszewski, senior reporter, news
When screen time is over, there’s no building toy, not even Lego, that holds my kid’s attention better than GraviTrax. It lets you build complicated marble runs using myriad components including ramps, risers, bridges, switches, and launchers — and that’s just the starter set. There are several expansions that add motorized components that can keep marbles rolling indefinitely if assembled correctly, and it’s one of those toys that adults will enjoy just as much as kids do. For younger kids, there’s also now a GraviTrax Junior line with simpler components and larger marbles. – Andrew Liszewski, senior reporter, news
Over the holidays, we gave several tech-y gifts to our four-year-old daughter, including walkie-talkies, a camera, and a piano (really for me). But the gift she loves the most is the MUID Benson Lying Flat Duck Night Light, which is exactly as it’s titled: a lamp in the shape of a duck lying flat. But this lamp is squishy, entertaining to look at, and has an option to stay on for only 30 minutes. This is perfect for us, and for her, because it gives her some autonomy at bedtime when we turn off the other lights, letting her have 30 minutes to draw or play in her bed before going to sleep. – William Joel, senior creative director
My toddler is obsessed with music and is always bopping along to whatever is playing. My husband bought her a Yoto Player, which is like a speaker cube with dials that kids can adjust on their own to change volume and scroll through songs, and it’s become our most-used kid gadget. There’s a low-fi screen on the front that depicts simple images to coincide with the song (or story) that’s playing. You can control it through an app, which, honestly, I never downloaded, or through analog cards that kids can pop into the Yoto slot themselves. Cards come in several different languages for a range of ages and applications, like bedtime songs or classic folktales, and it chills her out on car rides and when she’s getting ready for bed. – Kristen Radtke, creative director
Resale platform Poshmark is launching Smart List AI, an automated product listing tool for sellers, the company announced today.
The tool has been in beta testing since last fall and uses sellers’ photos to generate details like the type of product, the size and brand, the style and color, and other details, as well as a title and description for the listing. Like other platforms that have introduced AI listing tools, the promise is that Smart List AI will streamline the process for sellers and cut down on the amount of time it takes to list items. Sellers will still input the price of their item, Manish Chandra, founder and CEO of Poshmark, told The Verge via email.
Other platforms like eBay have also introduced AI-generated listings that fill in details like product titles and descriptions. And like eBay’s feature, Poshmark’s tool will depend on sellers actually confirming that the information is correct — in the secondhand market, accurate details for things like the brand, size, and condition of an item are essential.
In my experience on eBay, at least, AI-generated descriptions are not very helpful; they mostly read like marketing fluff. What’s actually helpful is a detailed list of any flaws, wear and tear, or precise measurements — things that an AI listing tool wouldn’t be able to generate.
Retailers, including Shopify and Amazon, have pushed generative AI shopping products on both the merchant and consumer ends. And whether its AI-generated product listings or entire garments “created” using AI, the aim is often the same: to increase the scale of production, but not necessarily improve the quality.
Smart List AI will be available to Poshmark users on iOS in February.
Mark Zuckerberg is planning some big changes for Facebook in a bid to make using the platform feel like it did “back in the day.” During the company’s fourth-quarter earnings call on Wednesday, Zuck said that Facebook is an important part of its more than three billion monthly users’ lives and that he’s working on making it “way more culturally influential than it is today.”
He didn’t divulge any specific changes, only that he’s “excited this year to get back to some OG Facebook,” and that the overhaul will be a significant focus for the company over the next 12 months. Zuckerberg also suggested that Meta may have to sacrifice its short-term business results in order to focus on product changes needed to improve Facebook.
“I think that’s sort of a fun and interesting goal that will take our product development in some interesting directions that we maybe haven’t focused as much on in the last several years,” Zuckerberg said during the earnings call Q&A. “I think it’s an investment area and something I’m going to spend some time on.”
Recapturing the OG Facebook vibe could mean several things. Many people like myself will think back to the era before it was overtaken by our boomer and Gen X relatives — online spaces are less fun with your family spectating. But let’s not forget that the original Facebook, known as Facemash, was a site Zuckerberg created to nonconsensually rank his female classmates at Harvard by attractiveness. I sure hope this isn’t the “masculine energy” he thinks we need more of, given he reportedly blamed former chief operating officer Sheryl Sandberg for making inclusivity changes at the company.
Zuckerberg’s comments about old-school Facebook don’t necessarily mean he actually wants to strip modern features out of it. He celebrated that the time users spent watching videos on Facebook was up by “double-digit rates year-over-year” earlier in the earnings call, and said he expects Reels on Instagram and Facebook to keep growing “regardless” of what happens with TikTok.
LG is adding Google Cast support to its range of hotel TVs, making it easier to stream content from your phone or other devices. Cast joins existing support for Apple Airplay, and LG is the first manufacturer to offer both options on its hotel TVs.
If you stay in a hotel room with a compatible LG TV, you should be able to scan a QR code with your phone and then start streaming video or music directly to the screen from a user interface you’re already comfortable with. That way you won’t have to enter any personal information or login details into the TV itself. LG says it uses network isolation to maintain privacy between hotel rooms, and that your connection will be maintained for the duration of your stay but severed upon checkout.
Some hotels, including Hyatt, already offer Cast through customized hospitality Chromecast hardware. LG has instead included casting directly into the TV, with no additional hardware required.
“Many hotel guests nowadays wish to stream their favorite content from a personal device to their in-room TV,” said Paik Ki-mun, head of LG’s information display business. “With LG Hotel TVs integrating Google Cast, they can enjoy a secure and convenient content-streaming experience, just like they’re used to at home.”
LG says Cast should work on any of its TVs running webOS23 or later. From its consumer line-up, that covers all of the company’s models from 2023 onwards, and a selection of 2022 models via updates — so you’ll still have to hope that your next hotel gave its TVs a recent upgrade.
Here’s what the certification logo looks like, which may be used on book covers, spines, and promotional materials.
The Authors Guild — one of the largest associations of writers in the US — has launched a new project that allows authors to certify that their book was written by a human, and not generated by artificial intelligence.
Certification is currently restricted to Authors Guild members and books penned by a single writer, but will expand “in the future” to include books by non-Guild members and multiple authors. Books and other works must be almost entirely written by humans to qualify for a Human Authored mark, with minor exceptions to accommodate things like AI-powered grammar and spell-check applications.
“The Human Authored initiative isn’t about rejecting technology — it’s about creating transparency, acknowledging the reader’s desire for human connection, and celebrating the uniquely human elements of storytelling,” guild CEO Mary Rasenberger said in a statement on Wednesday. “Authors can still qualify if they use AI as a tool for spell-checking or research, but the certification connotes that the literary expression itself, with the unique human voice that every author brings to their writing, emanated from the human intellect.”
Vodafone has made what it calls “the world’s first” satellite video call using a standard smartphone, in a test of a system it says will provide mobile broadband service to 4G and 5G phones without dedicated satellite hardware. The service, using satellites from SpaceX rival AST SpaceMobile, is expected to launch in Europe before the end of 2025.
AT&T and Verizon have also cut deals with Texas-based AST SpaceMobile to provide satellite-to-smartphone services across the US. The company has received FCC approval to begin testing its US-based coverage for AT&T this spring, meaning a full rollout is likely to lag behind Europe.
Last year SpaceX demonstrated its own Starlink satellite-based video call between two US-based employees using “unmodified mobile phones.” That means Vodafone’s effort may not quite be a world-first, though unlike SpaceX it made the call from a remote area that apparently has no existing cell service.
For Vodafone’s demonstration, an engineer (and, charmingly, his dog) made the call to Vodafone CEO Margherita Della Valle from an unnamed “remote mountain region” of Wales which the company says has never had mobile coverage before. The quality isn’t pristine — the video is choppy, with noticeable lag — but the call, which lasts about 45 seconds, does seem stable.
Vodafone partnered with AST SpaceMobile for the satellite service, using the five low Earth orbit BlueBird satellites the company has launched so far. The signal is routed through a space-to-land gateway which connects the satellites to Vodafone’s terrestrial network. Vodafone sees it as a “complementary” technology that can plug the gaps in its existing mobile network, covering remote areas including mountains and out at sea. The direct-to-smartphone satellite service is expected to “close the last remaining coverage gaps” in Europe in 2026.
AST SpaceMobile’s satellite system connects using standard 4G and 5G signals, so there’s no requirement to use a phone with dedicated satellite components. Beyond video calling, Vodafone says it offers a “full mobile broadband experience,” with peak speeds of up to 120 megabits per second, that “goes beyond other low Earth orbit satellite constellations which have so far only facilitated text messaging.”
Satellite connectivity is already available on certain iPhones and Google Pixel phones that include specific modem components, but is mostly limited to emergency alerts, location-sharing, and SMS messages. T-Mobile is beta testing its own US network in partnership with the Starlink Direct-to-Cell service that will also work with standard smartphones, though this will be limited to text messages at first, with calls and data to be added in the future.
“Vodafone’s job is to get everyone connected, no matter where they are,” says Della Valle. “We are bringing customers the best network and connecting people who have never had access to mobile communications before. This will help to close the digital divide, supporting people from all corners of Europe to keep in touch with family and friends, or work, as well as ensuring reliable rural connectivity in an emergency.”
There’s been no announcement yet about pricing for the service.
That was CEO Mark Zuckerberg’s message to investors during his company’s fourth-quarter earnings call on Wednesday. During the Q&A portion of the call with Wall Street analysts, Zuckerberg fielded multiple questions about DeepSeek’s impressive AI models and what the implications are for Meta’s AI strategy. He said that what DeepSeek was able to accomplish with relatively little money has “only strengthened our conviction that this is the right thing to be focused on.”
Zuckerberg noted that “there’s a number of novel things they did we’re still digesting” and that Meta plans to implement DeepSeek’s “advancements” into Llama. DeepSeek caused a massive sell-off in AI stocks due to fears that models will no longer need as much computing power. Zuckerberg tried to dispel concerns that the billions of dollars he’s spending on GPUs will go to waste: “I continue to think that investing very heavily in CapEx and infra is going to be a strategic advantage over time.”
His argument is in line with the growing consensus that computing resources will move from the training phase of AI development towards helping models better “reason.” In Zuckerberg’s own words, this “doesn’t mean you need less compute” because you can “apply more compute at inference time in order to generate a higher level of intelligence and a higher quality of service.” Meta is gearing up to release Llama 4 with multimodal and “agentic” capabilities in the coming months, according to Zuckerberg. He expects Meta’s AI assistant to reach one billion users this year.
He also took a thinly veiled jab at OpenAI, Anthropic, and other unprofitable startups by noting that Meta has a “strong business model” to support the roughly $60 billion it will spend on AI this year versus “others who don’t necessarily have business models to support it on a sustainable basis.”
Zuckerberg also made sure to praise President Donald Trump. “We now have a US administration that is proud of our leading companies, prioritizes American technology winning,” and “will defend our values and interests abroad,” he said. Moments before the earnings call started, news broke that Meta is paying Trump $25 million to settle a lawsuit he brought against the company for banning his account after the January 6th insurrection. (The vast majority of the money is going to pay for Trump’s presidential library.)
Meanwhile, Meta is a cash-printing machine. Revenue for the fourth quarter of 2024 was $48.39 billion — a 22-percent increase from the year-ago period — while net profit was a staggering $20.8 billion (up 43-percent from a year before). During the earnings call, CFO Susan Li said that Meta hasn’t “seen any noticeable impact” from its content policy changes on ad spending. 3.35 billion people used at least one of Meta’s apps daily in the fourth quarter — a 5-percent increase from the year-ago period.
