Photo by The Verge Nvidia plans to release a native GeForce Now app for Steam Deck “later this year,” according to a blog post. It’s already relatively straightforward to get Nvidia’s cloud gaming service set up on Steam Deck thanks to a special script from Nvidia, but a native app should be easier to install and will support up to 4K resolution and 60 fps with HDR when connected to a TV.Nvidia also plans to bring GeForce Now to some major VR headsets later this month, including the Apple Vision Pro, Meta Quest 3 and 3S, and Pico “virtual- and mixed-reality devices.” When GeForce Now version 2.0.70 is available, people using those headsets will be able to access an “extensive library of games” they can stream by visiting play.geforcenow.com in their browser.The company also says that two major titles from Microsoft will be available on GeForce Now when they come out this year: Avowed, which launches February 18th, and DOOM: The Dark Ages, which is set to be available sometime this year.

When it comes to tools you should use on your PC, Task Manager is one of the more useful tools for power users on Windows 11. It's quite common to use it for basic functionality like ending processes that refuse to close, but aside from that, there is a lot more it can do, too. If you want even more functionality, though, then you should try out an alternative like Task Manager DeLuxe.

The Philips Hue app already offers a lot of customizable lighting effects, but soon you’ll be able to created personalized ones with the help of AI. | Image: Signify Philips Hue users will soon be able to create their own AI-powered lighting scenes, the company announced on Tuesday.Philips Hue’s new generative AI lighting assistant lets users create personalized lighting scenes based on occasion, mood, or style. They can do so by typing commands like “Give me a scene for a garden party” into the Hue app and also by using their voice. The assistant will then either recommend scenes from the existing Philips Hue gallery or create a new one entirely. In addition to creating scenes, the AI assistant can also be used to adjust the settings of existing lights, including brightness levels and color. Image: Philips Hue The assistant will recommend scenes based on prompts. The news comes after Hueblog reported that Philips Hue was working on a generative AI feature in October. According to Hueblog, the “AI Playground” lets users create lighting scenes based on typed text like “Create a scene for a children’s birthday party,” but there was no mention of voice commands.Philips Hue isn’t the first to offer an AI-powered lighting assistant. Govee’s AI Lighting Bot generates similar dynamic lighting effects and also responds to voice prompts. Meanwhile, Nanoleaf’s “Magic Scenes” feature creates similar scenes, too, but only responds to typed short phrases like “tropical paradise.”Philips Hue says the AI assistant will be compatible with all of its lights. The company hasn’t divulged details about its exact availability, but stated it’ll start rolling out in the first quarter of 2025.The AI-powered lighting assistant isn’t the only piece of news Philips Hue announced on Tuesday. The Philips Hue Sync TV app now supports LG TVs, while the color-changing Philips Hue Datura ceiling light is now available in the US starting at $299.99. The frameless ceiling panel offers a pair of individually controllable lights and customizable color effects.Philips Hue also announced new smart home security features, including smoke alarm sound detection and the ability to stream live camera feeds to Amazon Alexa and the Google Nest Hub. Users can also now check out live camera views or schedule system settings straight from their devices.

CES 2025 is well underway, and one surprise that came from the show was confirmation from Valve that the company wants SteamOS to run on everything. It's been kind of expected for a while, but a direct confirmation of it along with a promise for more devices to come with it isn't necessarily something most people had on their bingo cards for the show. What it means is that you'll be able to install SteamOS on your PC or your gaming handheld, without a need to worry about dealing with Windows.

It's easy to get caught up with all the hot new products in CES 2025, but not much attention goes towards all the progress and collaboration that goes on behind the scenes to make said products happen. If you want a peek into the inner workings of your favorite devices, Google and MediaTek took to the stage to reveal how their partnership is going. By the sounds of things, the two are going to make smart homes even better in the coming new year.

The 2025 Asus ROG Flow Z13. Asus has a new version of its Surface Pro-like gaming tablet for CES, and it’s making some sizable changes both inside and out. The Asus ROG Flow Z13 for 2025 is once again a slightly chunky, almost-half-inch-thick, 13-inch tablet with a built-in kickstand, magnetic keyboard cover, a bunch of ports, and a clear window on its rear with RGB lighting to show off its innards.That fun glass window is now larger, with a direct view of the motherboard, but the biggest change for the ROG Flow Z13 is its switch to integrated graphics. That may seem like a step backward for a gaming-focused tablet since gamers covet dedicated GPUs, but Asus is outfitting it with AMD’s powerful new “Strix Halo” processor. The ROG Flow Z13 can be configured with the Ryzen AI Max 390 for $1,999.99 or the Ryzen AI Max Plus 395 for $2,199.99. The top-end model with the Max Plus 395 has 16 CPU cores and 40 graphics cores, while the base-model Ryzen AI Max Plus 390 (curse these names) has 12 CPU cores and 32 graphics cores. The Z13 utilizes a redesigned stainless steel vapor chamber for cooling these graphics-heavy chips, which are capable of 120W TDP. All that power in the Z13 is responsible for driving a 13-inch, 2560 x 1600 touchscreen display with a speedy 180Hz refresh rate (up from 165Hz on the last-gen model), which you don’t often find in laptops and tablets of this size. For ports, it’s got two USB 4, one USB-A 3.2 Gen 2, HDMI 2.1, a microSD card slot capable of UHS-II speeds, and a 3.5mm combination headphone / mic jack. It’s also got a 5-megapixel front-facing webcam and 13-megapixel rear-facing camera — so you can flash your RGB as you awkwardly take tablet photos in public. It also has Wi-Fi 7.The Z13 supports USB-C Power Delivery for charging, but that won’t be powerful enough to allow its full performance under load. Instead, it comes with a 200W power adapter that uses Asus’ proprietary and reversible slim power jack — like on its recent laptops.Other quality-of-life improvements for the Z13 include a new detachable keyboard with larger keycaps and a more generously sized touchpad. And on its right side, beside the power button and volume rocker, is a new “ScreenXpert” button that summons a Command-Center-like widget that includes multiple-display window management controls, quick access to operating modes like Turbo mode or Silent mode, and other settings like muting your mic. It’s primarily there to help control things while in tablet mode since the keyboard contains shortcuts for most of these functions.I got a quick glimpse of the new ROG Flow Z13 at a preview event, and Asus sent me a preproduction model right before CES to get a little bit of hands-on time. It’s what I’m writing this post on right now, and boy do I appreciate the updates to this keyboard cover. The 1.7mm key travel and bigger touchpad go a long way toward getting work done. While the Ryzen 395 chip has the potential to be power-hungry, the battery life on the Z13 shows some promise. Asus is only claiming 10 hours of battery life, and I did manage to get through a full eight-plus-hour workday of Chrome tabs, streaming music (though the speakers seem kind of bad at first listen), and writing across multiple virtual desktops the day before flying to CES — with pretty much no issues.I definitely prefer a proper laptop to a tablet with a kickstand and keyboard cover, but being able to remove the keyboard deck for a little more flexibility and comfort when it’s time to fire up a game is pretty slick. I tried out a little Helldivers 2 on the Flow Z13, and it performed quite well, especially for a tablet. Set to the Z13’s native 2.5K resolution, in-game render scale on Ultra Quality, and texture details on medium, I saw 60fps or just slightly under, and it looked really nice. If I bumped it down from Ultra Quality to Quality scaling, it jumped up to an even smoother 80fps. This was, of course, while the tablet was plugged in and its fans were blasting on Turbo mode. Diving in again while unplugged dropped the Ultra Quality render scale performance down to the 45 to 50fps range since playing on battery limits you to Performance mode instead of Turbo.This is preproduction hardware, but so far, it’s pretty impressive for integrated graphics. AMD’s new chip might have something special here for thin and light devices, but since it lacks Thunderbolt 5, that means the Flow Z13 can’t use the full GPU bandwidth of Asus’ new XG Mobile eGPU. (Previous models could use the older XG Mobile via its proprietary connector.) But of course, that would make this somewhat portable PC gaming solution a little less portable, and the new XG Mobile costs about as much as the Flow Z13 itself.But does a gaming tablet make much sense in 2025 when portable PC gaming is being so adequately served by the Steam Deck and a bunch of other dedicated handhelds? We’ll have to see how a production model of the ROG Flow Z13 fares when it launches sometime in February.Photography by Antonio G. Di Benedetto / The Verge

Nvidia has spent the last two years showing off its digital “ACE” characters that can have in-game conversations with you using generative AI. But at CES 2025, the company is taking the ACE characters a step further by showing how they can be “autonomous game characters” — including, sometime this year, a teammate to help you get a chicken dinner in PUBG. Nvidia says that ACE characters can use AI to “perceive, plan, and act like human players,” per a blog post. “Powered by generative AI, ACE will enable living, dynamic game worlds with companions that comprehend and support player goals, and enemies that adapt dynamically to player tactics.” The characters are powered by “small language models (SLMs)” that are “capable of planning at human-like frequencies required for realistic decision making” as well as “multi-modal SLMs for vision and audio that allow AI characters to hear audio cues and perceive their environment.”As for how that will work in PUBG, you’ll be able to team up with the “PUBG Ally,” which Nvidia and PUBG publisher Krafton are calling the “world’s first Co-Playable Character (CPC).” The Ally will be able to “communicate using game-specific lingo, provide real-time strategic recommendations, find and share loot, drive vehicles, and fight other human players using the game’s extensive arsenal of weapons,” Nvidia says. Basically, it sounds like an AI teammate you can talk with natural language who’s supposed to be as capable as a human. And a video shows the Ally indeed helping a player find specific loot, bringing over a vehicle, and attempting to flank opposing players. But the video is heavily edited and isn’t live, so I’m skeptical that the Ally will work as well as is being shown here.AI characters built with ACE are coming to other games, too. Naraka: Bladepoint Mobile PC Version will get a “a local inference AI Teammate feature” in March 2025, while Naraka: Bladepoint on PC will get the feature “later in 2025,” according to Nvidia’s blog post. “AI Teammates powered by NVIDIA ACE can join your party, battling alongside you, finding you specific items that you need, swapping gear, offering suggestions on skills to unlock, and making plays that’ll help you achieve victory.”Krafton’s upcoming life simulation game called inZOI will also get CPCs called “Smart Zoi.” And Nvidia says that ACE characters can be bosses, too, and they’ll be used for boss encounters in Wemade Next’s MIR5.

CES 2025 is an exciting time for all tech enthusiasts, but it's especially amazing for gamers. It gives us an insight into how 2025 will look regarding hardware and what CPUs and GPUs will power the newest triple-A games in our PCs and laptops. Now, Asus ROG is showing off all of the new devices you can expect to see released in the coming year, and there are some pretty mighty rigs that are worth your time.

There have been plenty of keyboards shown off during CES 2025, but not all of them look as classy as what Keychron showed off during the event. If you want a keyboard that looks the part but you don't want to sacrifice performance, the Keychron K2 HE is likely what you're looking for. Best of all, you can grab one right now.

The 2025 Asus XG Mobile, now with standard Thunderbolt 5 instead of a proprietary connector. | Photo by Antonio G. Di Benedetto / The Verge This smoky black translucent box isn’t a gaming PC — instead, it might be the most powerful single-cable portable docking station ever conceived. When you plug your laptop or handheld into the just-announced 2025 Asus XG Mobile, it promises to add the power of Nvidia’s top-flight GeForce RTX 5090 mobile chip, and up to 140 watts of electricity, and two monitors, and a USB and SD-card-reading hub, and 5Gbps ethernet simultaneously.That’s because it’s the world’s first* Thunderbolt 5 external graphics card and one of the first Thunderbolt 5 docks, using the new 80 gigabit per second bidirectional link to do more things with a single cable than we’ve ever seen before. Photo by Antonio G. Di Benedetto / The Verge The 2025 XG Mobile’s ports — and a standard AC power connector, because the power supply lives inside. And if you’re keeping score, I’m pretty sure it’s also the first standards-based portable eGPU with an Nvidia graphics chip. While Asus’ last-gen XG Mobile also boasted up to an Nvidia 4090, you could only tap into that power with a proprietary port found only on a few Asus devices. (Its USB4 and Oculink rivals have mostly featured the AMD Radeon 7600M XT.) None of that makes it the most powerful eGPU out there, as I currently have no performance figures from Asus, and you can definitely go further with bigger docks that can fit desktop graphics cards rather than mobile GPUs. But Asus rep Anthony Spence tells me that the Thunderbolt 5 link does give you up to 64Gbps of bandwidth for its Nvidia graphics — more than USB4 and tied with Oculink — and I’m wowed that Asus managed to fit all this and a 350W power supply (no external brick!) into a sub-2.2-pound package with a fold-out kickstand.Asus says it’s even 25 percent lighter and 18 percent smaller than the previous proprietary model. It’s got HDMI 2.1 and DisplayPort 2.1 for video output and a pair of 10Gbps USB-A ports, in case you’re wondering. Image: Asus Note that it comes with a little vertical stand, too. When it arrives later in Q1, it won’t come cheap. Spence says the top-tier XG Mobile with an RTX 5090 laptop chip will cost $2199.99 — meaning you could almost certainly cobble together a more powerful (but stationary) solution yourself. That said, Asus does plan to sell a lower-end $1,199.99 version with Nvidia’s mobile RTX 5070 Ti. Again, you’re paying for compact power here rather than maximum bang for the buck. Photo by Antonio G. Di Benedetto / The Verge Yes, that Asus ROG logo is light-up, programmable RGB using the company’s Aura Sync. You can also make out the top-mounted SD card receptacle. While it should work with any Thunderbolt 4 or USB4 laptop or handheld, including Asus’ own ROG Ally X, you’ll likely want the still-rare Thunderbolt 5 to get the full GPU bandwidth here. Finding a Thunderbolt 5 computer that doesn’t already have a powerful discrete GPU might be tough, but perhaps some of 2025’s thin-and-light laptops will seize this opportunity to double as potent travel desktops. *We are aware of one possible Thunderbolt 5 eGPU enclosure, to house a desktop graphics card, but that WinStar has barely even been detailed yet.

Plenty of lights to go around. | Photo by Antonio G. Di Benedetto / The Verge Following a teaser last month, Asus’ latest ROG Strix Scar gaming laptops have arrived and they’re leaning all the way into the gamer aesthetic. The 2025 Scar 16 and 18 come with RGB lights all the way around the bottom of the chassis as well as a user-programmable LED dot-matrix display on the lid, as seen on other ROG devices like Asus’ gaming phones. Beneath the flashy exterior, the Scar 16 and 18 can be maxed out with an Intel Core Ultra 9 275HX processor and Nvidia GeForce RTX 5090 GPU. It can also be configured with up to 64GB of DDR5-5600 RAM and a 2TB PCIe Gen4 SSD. The ROG Nebula HDR display comprises a 16:10 2.5K Mini LED panel with a peak brightness of 1,200 nits and a 240Hz refresh rate. There are two Thunderbolt 5 ports included, and the design allows for easy access to the bottom panel for component upgrades. The Strix Scar 16 and 18 have all the cooling tech you’d expect from a gaming laptop of this caliber, including an end-to-end vapor chamber and sandwiched heatsink. Combined with the Conductonaut Extreme liquid metal treatment on the GPU and CPU, Asus claims that it can keep fan noise levels to a library-like 45dB, even during extended gaming sessions.On top of all that, the ROG Strix Scar comes with the aforementioned light show. Asus calls it AniMe Vision, and you can customize it to display personalized animations and sync it with any other AniMe Vision devices you own. Download some prebaked artwork or cook up your own using Asus’ pixel editor — the choice is yours.The ROG Strix Scar starts at $2,599; Asus says its new gaming laptops will begin shipping in February.

Image: The Verge Google DeepMind is assembling a new team of artificial intelligence researchers to develop “world models” that can simulate physical environments. The initiative will be led by Tim Brooks, a former co-lead for OpenAI’s Sora project who joined DeepMind in October to work on Google’s video generation and world simulators.World models are a relatively new development within AI that could serve a variety of purposes, such as creating real-time interactive media environments for video games and movies, and realistic training scenarios for robots and other AI systems. It’s also part of Google’s push to achieve an artificial general intelligence system, or AGI, before its competitors.“DeepMind has ambitious plans to make massive generative models that simulate the world,” Brooks announced in an X post on Monday. Brooks included two open job listings for research engineers and scientists who will help to advance AI “world models” capable of simulating real-world scenarios by solving problems around training “at massive scale,” curating training data, and studying how they can be integrated with multimodal language models.DeepMind has ambitious plans to make massive generative models that simulate the world. I'm hiring for a new team with this mission. Come build with us!https://t.co/pqvALtAvLs https://t.co/vtwgeXl9Dl— Tim Brooks (@_tim_brooks) January 6, 2025“We believe scaling pretraining on video and multimodal data is on the critical path to artificial general intelligence,” DeepMind said in the job descriptions. “World models will power numerous domains, such as visual reasoning and simulation, planning for embodied agents, and real-time interactive entertainment.”The race to be the first to declare AGI is heating up, so Google’s focus here isn’t surprising. OpenAI CEO Sam Altman recently said that the company has cracked how to achieve the tech industry’s long-sought benchmark, and that autonomous AI agents may start to meaningfully join workforces this year.There are several competitors who already have a head start over DeepMind’s world modeling project, including Nvidia’s Cosmos platform for advancing physical AI, autonomous vehicle, and robot development; and the World Labs startup created by “the godmother of AI,” Fei-Fei Li.Introducing Genie 2 - our most capable large-scale foundation world model, which can generate a diverse array of consistent worlds, playable for up to a minute. We believe Genie 2 could unlock the next wave of capabilities for embodied agents . pic.twitter.com/AfL3EbOMeB— Jack Parker-Holder (@jparkerholder) December 4, 2024The new DeepMind team will work alongside existing Google AI projects including its flagship Gemini AI models, Veo video generator, and Genie — Google’s prior world model for simulating playable 3D environments in real-time.