A 3-inch touchscreen for swiping and zooming on roads and trails. Garmin’s $449.99 eTrex Touch is a rugged GPS navigator with preloaded maps for outdoor adventures. It features a 3-inch touchscreen for swiping and zooming in on roads and trails, and comes preloaded with TopoActive maps and the option to add an Outdoor Maps Plus plan ($4.99/month) for premium content like highly detailed elevation contours and satellite imagery, land boundaries, and wildlife areas.Its beefy girth betrays a 130-hour battery within, that can extend to about 650 hours in expedition mode for almost a full month of use before needing a charge.

Two Russian YouTubers have posted videos unboxing what appears to be an unannounced iPad Pro with an M5 chip. Rumors have indicated that Apple could launch an M5 iPad Pro as early as October, and these videos may be our first look at the actual product.Externally, this new M5 iPad Pro doesn’t appear to have major differences from the M4 iPad Pro released in May 2024. Based on what’s shown in the videos, the new device is still a big, thin tablet with a relatively unobtrusive bezel, and it once again looks to only have one camera on the back.The M5 iPad Pro is rumored to get a portrait front camera in addition to the landscape front camera available on the M4 model. In the video from one YouTuber, Wylsacom (who leaked an M4 MacBook Pro last year ahead of Apple’s official announcement), it’s difficult to see if this alleged new iPad has two front-facing cameras. The video from another YouTuber, Romancev768, shows a second sensor on the top of the iPad in portrait mode, but we haven’t seen anyone use it as a camera yet so it’s unclear what it may do.The iPads in the videos are also missing the “iPad Pro” text found on the back of the M4 model — the only things you can see on the back of the new tablet are the antenna line, the camera system, and the Apple logo.Wylsacom ran some Geekbench 6 tests on this alleged M5 iPad Pro and an M4 iPad Pro to compare benchmarks across the two models. For the CPU tests, the M5 iPad Pro got a single-core score of 4,133 and a multi-core score of 15,437, while the M4 iPad Pro got a single-core score of 3,718 and a multi-core score of 13,324. For the GPU tests, the M5 iPad Pro got a score of 74,568 and the M4 iPad Pro got a score of 55,702.Apple didn’t immediately reply to a request for comment. But if Apple is indeed going to launch an M5 iPad Pro sometime soon, we may not have to wait too much longer to see if the tablet in these videos are the real deal.Update, September 30th: Added details from a second video.

For EVs, it's Zero Hour. When the clock strikes midnight on Oct. 1, the $7,500 federal tax credit on EVs will expire, potentially turning affordable electric carriages into showroom pumpkins.Sales of electric cars are certain to dip. Real-world prices will rise. The only question is how much, and for how long, as President Donald Trump and his administration continue to go scorched-earth over climate change and kneecap support for electric cars and renewable energy.With one eye on the clock, consumers bought more new EVs in August 2025 than in any other month in US history: Sales reached 146,332 cars, up nearly 18 percent year over year, …Read the full story at The Verge.

Nothing’s Playground tool lets users design smartphone apps with AI. Nothing says it has just taken “the first step towards an AI-native operating system.” Not exactly. Playground, Nothing’s new app store filled with user-designed, AI-generated apps, is neither Nothing’s first step, nor is it an operating system — it’s built on Android. But the idea lands, and beneath the hype, it is genuinely exciting — a glimpse of smartphones that feel more personal and adaptable, with a clear vision for where this might go in the future. The bold proclamation of an AI-native OS centers on the launch of Essential, which Nothing says brings us closer to a world “where devices adapt to people, not the other way around.” In plain speak, Essential is Nothing’s “name for all our AI-related products,” founder and CEO Carl Pei tells The Verge. Despite the London startup claiming to be unveiling Essential today, the line actually has several products live already, including an AI search tool and Essential Space, which organizes voice notes and images.What’s really new is Playground, and with it Essential Apps. Playground is basically Nothing’s take on an app store. Filling it will be apps designed in Essential Apps, an AI-powered tool that can build simple apps from written prompts alone. In short, users can design their own apps for their own needs. Nothing suggests a mood tracker that syncs with a music playlist, a receipt-to-expenses-form pipeline, or an app that can suggest what to wear based on your wardrobe and upcoming appointments.  It’s clear this is the early stages of what Pei imagines as a much larger ecosystem, making some of Nothing’s “first step” comments make more sense. At the moment, you can only design widgets, which obviously limits what they can do. Apps must also be created on a web platform with written inputs, at which point creators can install them on their own phones or add them to Playground for others to download. It’s available exclusively to Nothing phones, with the exception of the Phone 1, which Pei says no longer receives major updates so can’t install the new apps. Over time, Pei tells The Verge this process of app creation will become “more and more seamless,” taking place directly on the phone itself and perhaps by just speaking. Widgets are not the end game either, Pei says, forecasting more capable apps that take up the entire screen, comparable to the standard apps we have today. There are also the kernels of what could become a new creator economy, if the idea takes off. Playground lets users take and remix apps others have made, building and evolving in a similar fashion to the open source community does today. For someone who speaks a lot about the moats smartphone giants Apple and Google have built around their app stores,  Pei claims he is not thinking about how to monetize the store yet. “I think it needs to reach a certain level of scale before we can turn on whatever business model we have,” he says, adding that “YouTube could be a good example of where it could potentially go” for creators.   Eventually — and here’s where the AI-native OS idea comes in — Pei says phones could be more proactive, changing apps’ placement, or even suggesting apps based on how we use the phone. But even then, this is not an OS. It is an interface. Pei admits as much, leaning on semantics to skirt the issue: “I guess the word or the noun ‘OS’ could be interpreted in different ways.” Pei tells The Verge Nothing has no intention of abandoning Google’s Android, the system on which its products are currently built.  “We’re not touching the lower level code,” he says. “Android is a great platform to build upon, and Android already has a rich developer ecosystem,” Pei explains. People will still need to download apps like Instagram or Tiktok when starting up their phones, after all. It’s an exciting new feature and potentially a new and exciting creator economy. But there is a long way to go before Nothing can truly go it alone. In a departure from the hyped AI-powered gadgets like Rabbit’s R1 or OpenAI’s partnership with former Apple designer Jony Ive, Pei is not promising to do away with the smartphone just yet. It is clear Pei believes phones are here to stay when he asks himself, “When am I ready to say goodbye to apps?” and answers: “You know, not anytime soon.”

A frame from a Sora 2-generated video. OpenAI has a new version of the Sora AI video generator that it launched at the end of last year, and it’s arriving today alongside a new social video app, also called Sora, for iPhones. The currently invite-only app resembles TikTok with a feed of videos you can shuffle through. But instead of encouraging people to stitch together duets, it asks you to record short videos that anyone can spin into new AI-generated deepfakes — with your consent.In a briefing with reporters on Monday, employees called it the potential “ChatGPT moment for video generation.” The Sora app is currently only available to US and Canada users, with other countries set to follow, and when someone receives access, they also get four additional invites to share with friends.  There’s no word on when an Android version might be released.Sora users can give their friends — or, if they’re feeling bold, everyone — permission to create “cameos” with their own likeness using the new video model, which is dubbed Sora 2. The person whose likeness is being generated is a “co-owner” of that end result, OpenAI employees said, and they can delete it or revoke access to others at any time. Like TikTok, OpenAI’s Sora app allows you to interact with other videos and trends using a “Remix” feature, but it only allows for the generation of 10-second videos for now. OpenAI currently has a lot of restrictions on generating videos of public figures, but that may not be the case forever, according to the briefing with reporters. Currently, “public figures can’t be generated in Sora unless they’ve uploaded a cameo themselves and given consent for it to be used,” OpenAI wrote in a release accompanying the announcement. “The same applies to everyone: if you haven’t uploaded a cameo, your likeness can’t be used.” OpenAI employees also said it’s currently “impossible to generate” X-rated or “extreme” content via the platform.

Spotify CEO Daniel Ek has announced that he’s stepping down as CEO, two decades after founding the music streaming service. Ek will transition to an executive chairman role at the company on January 1st, 2026, with Spotify co-presidents Alex Norström and Gustav Söderström set to step in as co-CEOs.“As Executive Chairman, I will spend more of my time on the long arc: strategy, capital allocation, regulatory efforts, and the calls that will shape the next decade for Spotify,” Ek said in a statement. “Gustav and Alex will continue to report to me, and we will work closely together with our Board of Directors.”As co-presidents, Norström currently serves as Spotify’s chief business officer, while Söderström leads the company’s product and technology unit. Ek says his new executive chairman role reflects a European setup, and that he will retain a hands-on approach to shaping the future of the company.This story is developing…

As LLMs have grown in popularity, the ability to run them locally has also become somewhat sought after. And it's not always easy, as the raw power required for running a lot of language models isn't something you can typically find in an ordinary laptop, and you'd probably think about using discrete GPUs first.

Microsoft is rolling out its annual Windows 11 update (known as version 25H2) today. After testing the update in the Windows Insider Release Preview ring last month, version 25H2 is now starting to roll out to all Windows 11 users through Windows Update. The 25H2 update isn’t a huge one, so it won’t take long to install — Microsoft is delivering it as a small “enablement package.”“This means that it contains all the recently announced features in Windows 11, version 24H2,” explains John Cable, vice president of program management for Windows servicing and delivery. “Both versions use a shared code base and servicing branch, streamlining the update.”This shared code base also means Microsoft will continue to release new Windows features monthly to both 24H2 and 25H2 versions of Windows 11. Version 25H2 does have some improvements behind the scenes, though. “Version 25H2 includes significant advancements in build and runtime vulnerability detection, coupled with AI assisted secure coding,” Cable says. “We’ve designed version 25H2 to address and mitigate security threats in adherence to a robust security development lifecycle (SDL) policy and requirements.”The update is also a little leaner thanks to the removal of PowerShell 2.0 and the Windows Management Instrumentation command-line (WMIC).If you want to get 25H2 today, then all you need to do is ensure the “get the latest updates as soon as they’re available” option is toggled on in Windows Update. “If we detect that your device may have an issue, such as an application or driver incompatibility, we may put a safeguard hold in place and not offer the update until that issue is resolved,” Cable says. Microsoft will expand the rollout over the next few months and share known issues and device incompatibilities on its Windows release health hub.

A Raspberry Pi makes an excellent entry point into home server projects. While it won’t replace a full-blown server or NAS in raw power, it’s versatile enough to handle a surprising range of services. With a small footprint, low energy usage, and strong community support, the Pi can be the heart of a self-hosted setup that’s both fun and practical.

Just over three months ago, I managed to get a sneak peek at Opera Neon's AI tools. The build I got was a very basic, work-in-progress service, but it was still enough for me to remaster the game of Snake for modern audiences using the browser's "Make" mode. And while I was really impressed with what I managed even in a WIP state, I was a little sad that the browser wasn't ready for release at the time.

I've been exploring useful self-hosted tools that can help me organize my life — specifically, an app that can assist with digitizing physical documents. However, it can be challenging to find a solution that is both functional and user-friendly.

Non-Volatile Memory Express (NVMe) solid-state drives (SSDs) have transformed the way we store data. These drives offer speeds up to 15,000 MB/s (yes, that's 15 GB/s) that can load games, operating systems, and all other files in a few seconds. As well as raw speed, NVMe SSDs offer ultra-low latency and high throughput, making them an excellent choice for caching. SSD caching is useful when used in conjunction with slower storage, such as SATA SSDs and hard-disk drives (HDDs).