Succulents may be able to survive harsh conditions and human neglect, but these reliable potted plants deserve better. Here's how to properly care for succulents at home.
Nvidia’s incoming RTX 5090 has set tongues wagging again on the topic of how slim the next-gen flagship might just be.
VideoCardz noticed that Kopite7kimi, a regular on X (formerly Twitter) for GPU leaks, chipped in with another post that claims Nvidia’s version of the RTX 5090 (the Founders Edition) will be a two-slot graphics card.
While you might think that’s not exactly slim as such, it definitely is for a flagship GPU. The RTX 4090 spans three or more slots, eating up huge amounts of space inside the case of the PC it’s installed in.
The leaker further elaborates that the RTX 5090 will be a dual-fan card, so we’re looking at a two-slot cooler with two fans to keep the temperatures of the Blackwell flagship GPU in check.
Analysis: Efficiency in spades
For those who follow heavyweight GPUs, this rumor might be a difficult one to believe – and of course, we should exercise plenty of skepticism here (that’s true of any pre-release speculation).
However, this isn’t the first time we’ve heard about a slimmer RTX 5090, apparently due to a very different design approach Nvidia is taking with the flagship graphics card this time around. The source of that rumor mentioned a dual-slot cooling solution already (and that 32GB of VRAM might be the memory configuration for the 5090).
As we’ve noted before, with the RTX 5090 expected to make a big performance leap, this does beg the question of how much power it’ll use – and how good the cooling will have to be if it’s going to be a more compact solution in the next-gen flagship.
Kopite7kimi wouldn’t comment on the potential power usage of the RTX 5090, but does say: “I am sure the cooling design is more efficient.” This is certainly piquing our interest, as not only might the RTX 5090 forge ahead with gaming GPU performance, but it could also do that while being slimmed down – by a lot. This would fix one of our biggest issues with the RTX 4090 – namely, being able to fit the hulking thing into a PC case that isn’t the size of an aircraft hanger.
Does all this sound too good to be true? Well, to be fair it does, but we’ll just have to wait and see, though we may be in for a longer wait than is ideal, as the RTX 5080 might be the first Blackwell GPU to be released if the grapevine is correct. (That said, the RTX 5090 may not be far behind even in this eventuality).
We’ve already had our calendars marked for June 10 through 14 for Apple’s 2024 WWDC (Worldwide Developers Conference), but now we have official invites for the special event and a confirmed start time.
Apple just sent out invites to the press confirming the WWDC 2024 special event will kick off at 10am PT / 1pm ET / 6pm BST on June 10 (3am AEST, June 11). Unlike previous Apple event invites – including the May 7 “Let Loose” event – this WWDC invite lacks a tagline and has a simpler design. However, there is a return to classic Apple form as the event kicks off at the often-used 10am PT time.
(Image credit: Apple)
The invite features a glowing circle of blue, pink, purple, and yellowish-orange hues. Wondering why a circle? Well, it pretty much matches the overall shape of the main ring building at Apple Park, the company’s headquarters in Cupertino, CA. There is a thicker row, then a thinner one, and lastly, a thicker one with the center open, displaying the WWDC 24 logo. The colors match the original logo shared back in March 2024 by Apple.
This could be a nod to the in-person event for media and developers at Apple Park, but it might hint at an all-encompassing rollout for Apple’s rumored AI feature set.
(Image credit: Future/Lance Ulanoff)
In addition to the special event invites, Apple also shared a new main image for WWDC and updated its developer website. Both depict a version of the "Swift" logo with a holographic color scheme, but only the developer version has a tagline attached: "Coming in swiftly."
(Image credit: Apple)
Alongside the next generations of iOS, iPadOS, macOS, watchOS, tvOS, and visionOS, we expect Apple to debut its overarching approach to AI. We recently saw Apple add AI-infused and -powered features to Logic Pro for iPad and Mac, but WWDC 2024 will be different.
After OpenAI debuted ChatGPT-4o, Google showed off Project Astra, and Microsoft unveiled AI PCs, all eyes are now on Apple to deliver and impress on this front. It could be an LLM that runs on-device, with genuinely useful features for all the core applications, such as new photo editing tools or summarization tools across mail, calendar, and even phone. I hope we get transcription built into Voice Memos, at the minimum.
Specific to iOS 18, we expect more customization options for the home screen, smart summaries across nearly all applications, AI-generated emojis, and vastly improved searches. All of these were detailed in a fresh report from Bloomberg’s Mark Gurman, which sets out the new AI features to expect in iOS 18. We may also see Apple speak to its in-house-developed AI models and potentially LLMs that could integrate with a partner like OpenAI to bake ChatGPT into the operating system.
Whatever route it takes, we can expect the company to highlight the fact that a good portion of this will be processed securely and privately on-device, versus being sent off to the cloud. After all, the Neural Processing Units (NPUs) built into Apple’s A-series and M-series silicon chips should be more than capable of handling these tasks.
Considering Apple refreshed the MacBook Air with the M3 chip in March and just debuted the M4 chip in the new iPad Pro earlier this month (check out our iPad Pro 13 (2024) review for our verdict), it might be too early to see additional hardware. However, there is always the possibility that Apple will revealed souped-up versions of these chips in a Mac, such as the Mac Studio or Mac Pro.
If you're looking for a new iPhone, Apple Watch, iPad, or even a next-generation Vision Pro, those are likely all coming later. We could see Apple use WWDC 2024 as the event to announce the Vision Pro's international rollout, though.
(Image credit: Apple)
Either route, time will tell on all this, and at least we know when we should all be tuning in – 10am PT / 1pm ET on June 10, 2024, to watch Apple unveil it all from Apple Park. TechRadar will be on the ground, so stay tuned for full coverage. In the meantime, check out our guide on how to watch WWDC 2024 so you don't miss a thing.
My first Made In purchase won me over and now I'm eyeing the premium brand's other pots and pans as part of my Memorial Day deals shopping plan -- shelf space be damned. Here's why I think you'll love it, too.
Adjustable bases offer the most control over your sleeping experience. From special features to perks, here are the best adjustable bed bases you can get.
Snapseed lives to fight – or should I say edit – another day, following a recent update that was rolled out on the Play Store for Android devices, according to Android Police.
I'm thankful - the long time app, which is owned by Google and has over 100 million downloads, remains one of my favorite free photo editors with some seriously powerful tools you simply don't get with your Android device's camera app.
Snapseed's smart tools include a wide range of one-push looks and a neat 'Selective' feature that can make localized edits, for instance, brightening just your subject to make them pop.
You'll be hard-pressed to see exactly what is new in this recent update, though. There's an auspicious lack of AI-powered tools that are regularly at the fore today in Google's own Google Photos app, and the swathe of AI-powered apps cropping up for iOS and Android devices, such as Canva.
Since Google purchased Snapseed in 2012, updates have been few and far between. Major updates include 2017's redesign and a refresh in 2023 that made it compatible with Android 13.
And so while the latest Snapseed update provides solace to its regular users that the free editor is going nowhere for now, it feels like the writing is on the wall for one of the most loved free photo editing apps, given 2024's rival apps rely so heavily on AI.
With the latest update, it's fair to assume that Google is putting most of its efforts into developing its own Google Photos app, likely future AI-powered tools. What its plans are with Snapseed remains to be seen.
Perhaps it's nostalgia on my part, but I'm hoping I can still use my long-time favorite free editing app for years to come. It remains a powerful editor despite its lack of AI, as you can see in one of my photo edits above. Maybe Google will change the tide for the editing veteran and introduce new, powerful tools, but that seems unlikely. Just please don't retire it.
Storage is a major concern for business struggling to back up and manage all of the data they generate - a situation not helped by the arrival of generative AI which has massively disrupted traditional storage solutions.
Huawei’s answer to this problem is its OceanStor A800 NAS device. Originally unveiled last year, Huawei showed it off properly to assembled European journalists and partners at the recent Innovative Data Infrastructure (IDI) Forum 2024 in Berlin.
As we reported, the OceanStor A800 is capable of 24 million IOPS per controller enclosure, delivering ten-times the performance of existing storage, and ten-times the data mobility. It supports bandwidth in PB/s and 100 million IOPS, and, according to Huawei, offers 99.9999% data reliability.
Exclusive to OceanStor A800 NAS
For now, storage for the OceanStor A800 comes in the form of 30 and 50TB SSDs, but that will change when the company introduces a 128TB SSD next year. Huawei said its massive new internal drive, teased previously but officially unveiled at IDI Forum 2024, will “consume 88% less storage space and 92% less energy than the peer vendor’s SSDs when storing every one PB of data.” Exactly who the "peer vendor" is wasn't revealed.
Michael Fan, Huawei Data Storage marketing VP, told Blocks & Files these new SSDs “will use a mix of TLC and QLC flash to optimize capacity and drive lifetime. They have a Huawei-developed controller and come in U.2 and palm-size physical formats.” The site also noted that “The palm-size format is thinner than a U.2 SSD and could be either longer or shorter than that format. It appears to be similar to the EDSFF E3.S format but Huawei does not use EDSFF nomenclature.”
Although Huawei hasn’t said too much about the 128TB SSDs, or the proprietary technology behind them, it seems likely that they will be used exclusively in its OceanStor A800 NAS device for now, although it's possible it may become more widely available in the future.
(Image credit: Huawei )
More from TechRadar Pro
These are the best SSDs available to buy right now
If you’ve been curious about hybrid beds, now’s a great time to try one. Our sleep experts have tested the best hybrid mattresses available, and these are our current favorites.
When I first heard about Recall, I immediately buried my face in my hands. I never thought I'd see such a glaring target be created by Microsoft, never mind it being marketed as a feature.
If you haven't read about it yet, Recall is an AI feature coming to Windows 11 Copilot+ PCs. It's designed to let you go back in time on your computer by "taking images of your active screen every few seconds" and analyzing them with AI, according to Microsoft's Recall FAQs. If anyone other than you gets access to that Recall data, it could be disastrous.
Satya Nadella says Windows PCs will have a photographic memory feature called Recall that will remember and understand everything you do on your computer by taking constant screenshots pic.twitter.com/Gubi4DGHcsMay 20, 2024
See more
This might sound familiar, and that's because it's remarkably similar to the failed and shelved Timeline feature back on Windows 10. However, unlike Timeline, Recall doesn't just restore a version of your desktop files, it uses AI to take you back to that moment, even opening relevant apps.
What's the problem with Windows Recall?
On the surface, this sounds like a cool feature, but that paranoid privacy purist in the back of my mind is burying his face in a pillow and screaming. Imagine if almost everything you had done for the past three months was recorded for anyone with access to your computer to see. Well, if you use Recall, you won't have to imagine.
That might seem like an overreaction, but let me explain: Recall is taking screenshots every few seconds and storing them on your device. Adding encryption into the mix, that's an enormous amount of bloaty visual data that will show almost everything you've been doing on your computer during that period.
As Microsoft explains, "The default allocation for Recall on a device with 256 GB will be 25 GB, which can store approximately 3 months of snapshots. You can increase the storage allocation for Recall in your PC Settings. Old snapshots will be deleted once you use your allocated storage, allowing new ones to be stored."
This is worse than keylogging! Recall isn't just recording what you type, it's recording everything you're doing, with photo evidence, every three seconds.
This is worse than keylogging!
I say almost everything because Microsoft claims "Recall also does not take snapshots of certain kinds of content, including InPrivate web browsing sessions in Microsoft Edge. It treats material protected with digital rights management (DRM) similarly; like other Windows apps such as the Snipping Tool, Recall will not store DRM content." That's reassuring on the surface, but it's still far too vague for anyone to actually have any faith in it.
Will this only work on Microsoft Edge, or will it integrate with Chrome and Firefox too? If it only works with Edge, that feels like an egregious walling off of privacy for not using Microsoft's unpopular web browser.
But that's just the tip of the iceberg. Microsoft openly admits that Recall will be taking screenshots of your passwords and private data:
"Note that Recall does not perform content moderation. It will not hide information such as passwords or financial account numbers. That data may be in snapshots that are stored on your device, especially when sites do not follow standard internet protocols like cloaking password entry."
So, what you could have here is something that stores your passwords, your information, your account details, etc, and that is visible to anyone on your profile. If you only have one profile for your device, that means everyone with access to that PC will be able to see your Recall data.
Arguably, the worst part about this is that it will be on by default once you activate your device. Microsoft states:
On by default
"On Copilot+ PCs powered by a Snapdragon® X Series processor, you will see the Recall taskbar icon after you first activate your device. You can use that icon to open Recall’s settings and make choices about what snapshots Recall collects and stores on your device."
I think this is a bad idea. The decision should be made by the individual, and not by Windows. Having it immediately active just means that uninformed people may not be able to act upon this. In my eyes, it's akin to cookie tracking – it can be just as invasive. All of this makes me wonder whether it may hit a snag with consent under GDPR.
Is Microsoft making Recall secure?
In defense of Microsoft, I'd like it to be known that there was an attempt to make it secure. I don't think it was a very good one, but there was an attempt.
Microsoft states that "Recall snapshots are kept on Copilot+ PCs themselves, on the local hard disk, and are protected using data encryption on your device and (if you have Windows 11 Pro or an enterprise Windows 11 SKU) BitLocker." From the wording here, that looks like your snapshots will only be encrypted if you have Windows Pro or a business Windows code.
The omission of Windows Home users is horrifying. If this is the case, it leaves everyday people vulnerable if their devices are compromised. People shouldn't have to pay a premium and upgrade to protect their privacy on an operating system that's snapshotting their screen every few seconds.
People shouldn't have to pay a premium and upgrade to protect their privacy
The big question, though, is what kind of encryption is being used? I've been working with virtual private network (VPN) encryption for a while now, and just because something is "encrypted" doesn't mean it's safe. In fact, with developments in quantum computing, encryption is under threat, and even the best VPN services are having to come up with quantum-secure encryption methods. We've already seen that BitLocker can be cracked.
Another note in Microsoft's favor is that the data is stored locally and encrypted, rather than it being uploaded to a cloud server for Microsoft to access.
"Recall screenshots are only linked to a specific user profile and Recall does not share them with other users, make them available for Microsoft to view, or use them for targeting advertisements."
This means that, for now, Microsoft isn't peeking behind the curtain. But that doesn't guarantee that'll be the case forever. If Microsoft can legally find a way to make money out of this tool, my guess is that they'll try. For now, the push seems to be to persuade people to upgrade their OS.
If you're one of those households that has different profiles for each person on the family PC, you can claw back a little bit of privacy.
"Screenshots are only available to the person whose profile was used to sign in to the device. If two people share a device with different profiles they will not be able to access each other’s screenshots. If they use the same profile to sign-in to the device then they will share a screenshot history. Otherwise, Recall screenshots are not available to other users or accessed by other applications or services."
The problem is, that's only helpful if you password-protect your profile, and if someone sets parental controls on your profile, that could give them a backdoor.
What are the security risks with Recall?
You're probably thinking "so what?" So let me give you a few scenarios where this could be a problem:
You're using a public computer: let's say you do some online shopping or banking on a library computer. You didn't realize Recall was active, and now the person using the computer after you has just gone into the Recall archive to pull up all of your bank details, your address, and your passwords. It's like handing your house keys over to a burglar before telling them you're going on holiday for the week.
You're using a work laptop: we've all used a company computer for personal reasons, be it looking at social media on your lunch break, or simply running some errands because you don't have your own laptop. Now your boss, your IT team, and anyone with access to your device, can go through and see every three seconds of how you're using their equipment. They could use this to track your work output and see how productive you are, they could even read private messages you send to people.
You're using a family PC: if you've been using the household computer, and you don't have a password-protected profile, anyone could walk in and open up your Recall history. If you've been doing anything unsavory it's about to be obvious, even if you deleted that search history.
You get hacked or your laptop gets stolen: this one's pretty obvious, but if someone manages to hack into your device, the encryption won't matter. Similarly, if someone just steals your laptop and you don't have a secure password locking it, then a criminal (cyber or otherwise) can use Recall to pull the whole world out from under your feet.
There are so many problems that can arise just from someone accessing your Recall data. Using a password manager would become irrelevant if someone can see you typing in your master password, your private messages will be anything but, and there's no point in deleting your search history because Microsoft is keeping the receipts!
How to protect your privacy with Windows Recall
There are a few ways you can protect your privacy from Windows Recall, but the obvious, and most effective one will be to disable it outright. As the saying goes "an ounce of prevention is worth a pound of cure." You're better off not having this stuff stored on your device in the first place.
If, however, you want to use Recall, you're going to need to do the following:
Make an individual profile on your PC: this will prevent people from having shared access to your Recall data as long as you follow my next tip.
Password-protect your profile: not just your device, but your profile, too. Don't use a weak password, be serious. Use three memorable words with numbers and symbols, and no, don't set your password as "3-Memorable-worD5!"
Encrypt your Recall data: you may have to upgrade your OS or pay for BitLocker, but encryption is a non-negotiable. If someone gets past your password, you don't want them to have immediate, unchecked access to what you've been doing for the past three months.
Don't access sensitive data while Recall is on: if you're going to type in personal passwords or look at NSFW content, just turn it off. This is obviously going to be annoying and time-consuming, but it's far better than the alternative of having it all screenshotted.
Bottom line: Recall makes my skin crawl
Look, I've been a privacy advocate and researcher for years. I don't like the idea of anything tracking what we do. But this... this is something else. The risk that comes with Recall, the sheer devastation it could cause if your device gets hacked, the idea that Microsoft may be walling off privacy behind what I can only describe as a paywall. It sickens me.
There is so much opportunity for misuse with this feature. Security cannot be understated. Privacy cannot be bolted on. Taking screenshots of my device from the second I activate my device should not be a default option. Put the user in control of their privacy, and put the decision in their hands.
All of this just pushes me into the privacy-loving flippers of Linux.
Today's clever cybercriminals use AI to alter videos' audio content, creating convincing deepfake scams, a problem that could only grow, especially in an election year.
To try and combat such threats, McAfee has joined forces with Intel to launch an AI-powered Deepfake Detector. Previously known as Project Mockingbird, the tool looks to identify and notify users when a video's audio is likely AI generated or manipulated.
McAfee's Deepfake Detector employs state-of-art AI techniques, including transformer-based Deep Neural Network models, so will only run on systems with an Intel Core Ultra processor. McAfee says the NPU in AI PCs improves the app’s performance by up to 300% and allows local deepfake inference, thereby avoiding sending sensitive user data to the cloud.
No plans to bring it to older Intel PCs
"In a world where seeing is no longer believing, where AI-generated deepfakes have made it harder than ever to tell real from fake, consumers need the latest in AI to beat AI," said Steve Grobman, McAfee's Chief Technology Officer.
"Leveraging Intel's Core Ultra processor technology and its NPU, we can provide consumers with the most advanced and powerful generative AI deepfake detection capabilities, without compromising on performance or privacy. No need to send huge files to the cloud for analysis with detection performed locally, providing even greater peace of mind and control over your data."
Carla Rodriguez, Vice President and General Manager of Ecosystem Partner Enabling at Intel, added, “We're excited to collaborate with McAfee to help bring to market innovative solutions like McAfee Deepfake Detector which leverages the NPU in Intel Core Ultra processor-based PCs, driving detection of malicious and misleading deepfakes, while delivering optimal performance and privacy."
McAfee’s Deepfake Detector will soon be available for English video detection. It is expected to be expanded to additional languages in due course.
We asked McAfee if there were any plans to bring the tool to older PCs and were told by a spokesperson, “McAfee Deepfake Detector isn’t being brought to market on older Intel PCs. The capabilities described involve leveraging the NPU in Intel Core Ultra processors. While we are proud to partner with Intel, another innovative, consumer-focused technology leader focused on delivering solutions to address the very real needs of consumers today, this is not an exclusive partnership.” That's potentially good news for anyone not running an Intel PC.
LattePanda has unveiled its latest single board computer, the LattePanda Mu - and although it is barely larger than a business card at a mere 60mm x 69.6mm, the micro x86 compute module is equipped with an Intel N100 CPU with 4 cores.
LattePanda says the Mu, which can run both Windows and compatible Linux operating systems, is faster than the Raspberry Pi 5, and more configurable. End-users can design custom carrier boards based on specific requirements, with suggestions including a dual-Ethernet port board for a router and a SATA interface for a NAS solution.
The Mu comes with 8GB of memory with IBECC support, 64GB of eMMC 5.1 storage, and a range of expansion pin options including three HDMI/DisplayPort, eight USB 2.0 ports, up to four USB 3.2 ports, up to two SATA 3.0, and 64 expandable GPIOs.
Add your own GPU
The Mu also offers up to nine PCIe 3.0 lanes, allowing for the addition of dedicated graphics card, so the diminutive device can be put to more demanding computational tasks.
The LattePanda Mu's open-source carrier board designs and libraries can be found on the company’s dedicated GitHub repository, allowing developers to make quick modifications and refinements.
“LattePanda Mu is not just a micro x86 compute module; it is a breakthrough design solution that offers users unlimited customization possibilities," said company product manager WangBo.
With its powerful computing performance, GPU capabilities that surpass traditional options, and flexible thermal design power range, it becomes the ideal choice for various applications and performance needs. The customization options of the LattePanda Mu allow users to design carrier boards according to their requirements, providing perfect solutions for diverse application scenarios.”
The LattePanda Mu board retails for $139, while a kit including the Mu, Mu Lite Carrier and Mu Active Cooler is currently available for the reduced price of $190.
If you’re intrigued by the LattePanda Mu but aren’t quite ready to buy it yet, we’ve got a full review coming up shortly.
Cybersecurity researchers from Intego have discovered new variants of the dreaded Cuckoo malware that targets macOS users.
For those unfamiliar with the name, Cuckoo is an infostealer targeting Mac devices running both on Intel and ARM silicon.
Intego’s researchers now say they have found a new variant that was pretending to be Homebrew, a popular macOS software package manager. The attackers set up a fake landing page, seemingly identical to the authentic Homebrew page, which deployed the infostealer.
Poisoning Google Ads
In early May 2024, Mac security provider Kandji said the malware “queries for specific files associated with specific applications, in an attempt to gather as much information as possible from the system." Apparently, Cuckoo was looking for hardware information, currently running processes, and installed applications.
Among its key features are the ability to take screenshots, harvest data from iCloud Keychains, Apple notes, web browsers, different apps (Discord, Telegram, Steam, and more), and grab cryptocurrency wallet data.
The threat was being distributed via fake software, a program claiming to be able to rip music from streaming services into .MP3 files.
While setting up a fake website is easy, getting people to visit it is infinitely harder. Intego believes that to get people to visit the website, the attackers engaged in Google Ads poisoning, obtaining access to Google Ads accounts with cleared and running campaigns, and modifying them (or running new campaigns) to generate traffic.
“We recommend that consumers get out of the habit of “just Google it” to find legitimate sites,” the researchers said. “Such habits often include clicking on the first link without giving it much thought, under the assumption that Google won’t lead them astray, and will give them the correct result right at the top. Malware makers know this, of course, and that’s why they’re paying Google for the number-one position.”
Instead of Googling popular websites, users are advised to type in the address themselves, or to bookmark the sites.
NVIDIA GeForce NOW’s claim to fame is its ability to let just about any computing device — a phone, Chromebook, or a smart TV — tap into its cloud gaming hardware to play the latest games with supreme performance. And since the service taps into libraries of games that players already own, there are no extra purchases.
With school about to end for the summer and plenty of extra time to game, NVIDIA has an exciting update in store for GeForce NOW. NVIDIA has acknowledged the prevalence of handheld gaming PCs like the Steam Deck and Asus ROG Ally, both excellent ways to tap into games on the go. However, these portables are making the most of integrated graphics, which conserves power and avoids heat, but there’s still plenty of uplift available through the use of more powerful hardware, which can better deliver the top-notch graphics, high resolutions, and smooth frame rates that gamers crave. So NVIDIA is giving them a helping hand. And since all the heavy lifting for GeForce NOW is done in the cloud, the handheld consoles get to relax, so you’ll still get great battery life.
GeForce NOW is geared up for easy use with the Steam Deck thanks to a new browser configuration. A quick beta installation brings a pre-configured Google Chrome installation to the Steam Deck, offering special settings that let players get up and running with GeForce NOW in the browser with ease. This setup allows the browser version of GeForce NOW to run in the Steam Deck’s Gaming Mode, allowing for simple navigation with the built-in gamepad. Beyond that, it also provides access to advanced graphics settings like DLSS and RTX real-time ray-tracing to substantially upgrade visuals. And this update isn’t just for the Steam Deck. It also allows for convenient navigation with the ASUS ROG ally, Logitech G Cloud, Lenovo Legion Go, MSI Claw, and Razer Edge.
With GeForce NOW, these portable handhelds can get a big boost from powerful computers in the cloud. You’ll be able to enjoy advanced graphics features like RTX real-time ray tracing and DLSS to enhance your game visuals. GeForce NOW supports high resolutions and frame rates up to 240 fps, letting you max out the gaming handhelds.
When you’re not on the move, it’s easy to pick up where you left off with GeForce NOW, since you can also run it on other devices. You can switch to a PC, Mac, NVIDIA SHIELD TV, or any of the many devices that support GeForce NOW.
For the Steam Deck, GeForce NOW also allows you to tap into other game libraries outside of Steam, such as Epic Games, GOG.com, Ubisoft Connect, and Xbox. You can even run PC Game Pass games through GeForce NOW on the Steam Deck. This is a big flexibility upgrade. With access to over 1,900 games and counting and no need to install each, storage on these small handhelds becomes a non-issue when you want to try a new game, and so does the handhelds’ ability to actually run the games on their own hardware. So when the next big game comes along or a hot new indie, you can jump right in. GeForce NOW will let you step aboard Honkai: Star Rail, blast through the wasteland in the Fallout series, explore other realms in Baldur’s Gate 3, tackle darkness in Alan Wake 2, and dig through your game backlog.
GeForce NOW’s list of supported games is constantly expanding, and NVIDIA is regularly adding new features. To stay up to speed with all the latest from GeForce NOW, you can tune into the weekly GFN Thursdays blog, which covers all the new games coming, feature updates, game sales, and plenty more.
from TechRadar - All the latest technology news https://ift.tt/WTEXm1U
The LG B3 is ranked in our best TV list as the best cheap OLED display, thanks to its stunning picture quality, excellent gaming features, and affordable price. Just ahead of this year's Memorial Day sales event, I spotted the 65-inch model on sale for $1,199.99 (was $1,499.99). That's a new record-low price and a fantastic value for a big-screen OLED TV.
Released last year, the LG B3 delivers an exceptional picture, thanks to the OLED display coupled with LG's α7 AI Processor Gen6, which results in deep contrasts and rich colors. You're also getting Dolby Vision IQ and Dolby Atmos for superior sound, excellent gaming features, and smart capabilities with webOS 23.
The best part about the LG B3 is, by far, the price. It's not only LG's cheapest OLED TV, but it's also one of the most affordable OLED displays on the market. If you're looking for a premium display on a budget, you can't get much better than today's deal on LG's 65-inch B3 OLED TV at Best Buy.
Today's best budget OLED TV: LG's 65-inch B3
LG 65-inch B3 Series OLED TV:was $1,499.99 now $1,199.99 at Best Buy The LG B3 OLED TV not only features a gorgeous display (120Hz refresh rate, 8.3 million self-lit pixels) that provides rich contrast and color, but it also comes with Nvidia G-Sync, AMD FreeSync Premium, and VRR built-in along with four HDMI 2.1 ports, making it perfect for gaming. Today's deal from Best Buy brings the 65-inch model down to a record-low price of $1,199.99.View Deal
More of today's best OLED TV deals
LG C2 42-Inch 4K OLED Smart TV (2022):was $1,549.99 now $849 at Walmart The LG C2 OLED was rated as last year's best TV, and Walmart has this 42-inch model on sale for a fantastic price of $849. The gorgeous display is praised for its intense brightness and vivid colors in our LG C2 OLED review and packs an a9 Gen5 AI Processor, Dolby Atmos, and voice control - all for under $1,000. Please note that this specific model is sold by a third-party seller but is fulfilled by Walmart.
LG C4 48-inch OLED 4K TV:was $1,599.99 now $1,499.99 at Best Buy LG's all-new 48-inch C4 OLED TV is getting a first-time $100 discount, bringing the price down to $1,499.99. The C4 is a successor to the highly-rated LG C3, and we predict it will become one of this year's best OLED TVs. Upgrades include new gaming features, LG's latest Alpha 9 AI chip for improved performance and exceptional brightness.View Deal
LG C3 65-inch OLED TV (2023):was $2,499.99 now $1,599.99 at Best Buy
Best Buy has the best-selling 65-inch LG C3 OLED TV on sale for $1,599.99, which is only $100 more than the record-low price we briefly saw a few weeks ago. The stunning OLED TV features a brilliant picture with bright colors and powerful contrast thanks to LG's latest Alpha9 Gen6 chip. Plus, you get four HDMI 2.1 ports for next-gen consoles, a sleek, thin design, and an updated webOS experience - all for under $1,600, which is fantastic value for a premium OLED display.View Deal
Samsung 65-inch S90C Smart 4K OLED TV:was $2,599.99 now $1,599.99 at Best Buy The Samsung S90C OLED is TechRadar's best TV of the year, and Best Buy has the 65-inch model on sale for $1,599.99. Our Samsung S90C review awarded this TV five stars, praising its gorgeous picture, extensive gaming features, super slim design, and reasonable price - especially with today's $1,000 price cut.View Deal
Shop more TV offers with our list of the bestTV deals, and if you're looking for a more premium display, see the bestOLED TV deals. You can also look forward to discounts at theMemorial Day TV salesevent.
from TechRadar - All the latest technology news https://ift.tt/HgEx5d0
I didn't expect Google Glass to make a minor comeback at Google I/O 2024, but it did, thanks to Project Astra.
That's Google's name for a new prototype of AI agents, underpinned by the Gemini multimodal AI, that can make sense of video and speech inputs, and smartly react to what a person is effectively looking at and answer queries about it.
Described as a "universal AI" that can be "truly helpful in everyday life", Project Astra is designed to be proactive, teachable, and able to understand natural language. And in a video ,Google demonstrated this with a person using what looked like a Pixel 8 Pro with the Astra AI running on it.
By pointing the phone's camera at room, the person was able to ask Astra to "tell me when you see something that makes sound", to which the AI will flagged a speaker it can see within the camera's viewfinder. From there the person was able to ask what a certain part of the speaker was, with the AI replying that the part in question is a tweeter and handles high frequencies.
But Astra does a lot more: it can identify code on a monitor and explain what it does, and it can work out where someone is in a city and provide a description of that area. Heck, when promoted, it can even make an alliterative sentence around a set of crayons in a fashion that's a tad Dr Zeus-like.
It can can even recall where the user has left a pair of glasses, as the AI remembers where it saw them last. It was able to do the latter as AI is designed to encode video frames of what it's seen, combine that video with speech inputs and put it all together in a timeline of events, caching that information so it can recall it later at speed.
Then flipping over to a person wearing the Google Glass 'smart glasses', Astra could see that the person was looking at a diagram of a system on a whiteboard, and figure out where optimizations could be made when asked about them.
Such capabilities suddenly make Glass seem genuinely useful, rather than the slightly creepy and arguably dud device it was a handful of years ago; maybe we'll see Google return to the smart glasses arena after this.
Project Astra can do all of this thanks to using multimodal AI, which in simple terms is a mix of neural network models that can process data and inputs from multiple sources; think mixing information from cameras and microphones with knowledge the AI has already been trained on.
Google didn't say when Project Astra will make it into products, or even into the hands of developers, but Google's DeepMind CEO Demis Hassabis said that "some of these capabilities are coming to Google products, like the Gemini app, later this year." I'd be very surprised if that doesn't mean the Google Pixel 9, which we're expecting to arrive later this year.
Now it's worth bearing in mind that Project Astra was shown off in a very slick video, and the reality of such onboard AI agents is they can suffer from latency. But it's a promising look at how Google will likely integrate actually useful AI tools into its future products.
How often do you read through terms and conditions, EULA’s and privacy policies? Although we know we should scour the fine print, it’s something few of us ever bother to do, and certainly not fully.
Non-profit organization Tax Policy Associates wanted to prove how pointless these documents are, and so in February 2024 added a line to its privacy policy, offering a “bottle of good wine” to the first person who spotted the offer and got in touch.
After three months of nobody noticing the addition, the reward was finally found by someone who chanced upon it after looking at several examples of privacy policies online to get an idea of how to create their own.
Not the first time
The organization's head, Dan Neidle, shared the story on X and told the BBC it was "my childish protest that all businesses have to have a privacy policy and no one reads it. Every tiny coffee shop has to have a privacy policy on their website, it’s crazy. It’s money that’s being wasted."
In its coverage, which was the most read story on the site, the BBC pointed out that any company that holds personal data, “including small businesses and charities”, has to have a privacy policy under the UK's General Data Protection Regulation 2018 (GDPR).
This is actually the second time that Tax Policy Associates has made a sneaky addition to its privacy policy. The first time it took four months to be found. "We did it again to see if people were paying more attention and they’re not," Neidle told the BBC.
The writing in the firm's privacy policy has since been changed following the discovery and now says, "We know nobody reads this, because we added in February that we’d send a bottle of good wine to the first person to contact us, and it was only in May that we got a response."
If you're wondering what counts as a "good" bottle of wine in this instance, the answer, according to the BBC, is a Château de Sales 2013/14, Pomerol.
Our ongoing experiment into whether anyone reads website T&Cs continues. We put this in our terms back in February. Just got claimed. pic.twitter.com/N7k3weTuA9May 9, 2024
We tested the best indoor gardens to find the right model for any gardener. Whether you're growing microgreens or a whole salad, you'll find the perfect indoor gardening system.
Academic researchers from multiple universities recently discovered a new Spectre-like method of extracting secrets from modern Intel processors. However, Intel says that the original Spectre mitigation fixes these flaws, too.
A group of researchers from the University of California San Diego, Purdue University, UNC Chapel Hill, Georgia Institute of Technology, and Google, discovered that a feature in the branch predictor called the Path History Register (PHR) can be tricked to expose sensitive data.
Thus, they dubbed the vulnerability “Pathfinder”.
Extracting AES encryption keys
"Pathfinder allows attackers to read and manipulate key components of the branch predictor, enabling two main types of attacks: reconstructing program control flow history and launching high-resolution Spectre attacks," Hosein Yavarzadeh, the lead author of the paper, told The Hacker News.
"This includes extracting secret images from libraries like libjpeg and recovering encryption keys from AES through intermediate value extraction."
For those with shorter memory, Spectre was a side-channel attack that exploited branch prediction and speculative execution in processors, allowing attackers to read sensitive data in the memory.
PHR’s job is to keep a record of the last branches taken. It can be fooled to induce branch mispredictions and thus cause a victim program to run unintended code paths. As a result, sensitive data gets exposed.
In the research paper, the academics demonstrated extracting the secret AES encryption key, and leaking secret images during libjpeg image library processing.
Intel was tipped off in November last year, and released a security advisory addressing the findings, in April this year. In the advisory, Intel said that Pathfinder builds on Spectre v1, adding that the previously released mitigations address this problem, as well.
AMD’s silicon seems to be immune to Pathfinder, the researchers concluded.
Those interested in learning more can read the entire paper on this link.
BIG-IP Next Central Manager (NCM), a centralized management and orchestration platform for F5’s BIG-IP product family, was vulnerable to two major flaws which allowed malicious actors to take over its managed assets.
The bugs, which have since been patched, are described as an SQL injection vulnerability, and an OData injection vulnerability.
They are tracked as CVE-2024-26026 and CVE-2024-21793, and are found in the NCM API. By abusing these bugs, threat actors could run malicious SQL statements on vulnerable endpoints from a distance.
Thousands of potential victims
Cybersecurity firm Eclypsium found and reported the flaws, and the researchers also published a proof-of-concept exploit, which demonstrates how a rogue admin account, created by an attacker, remains invisible in the Next Central Manager, granting persistence on the vulnerable endpoint.
"The management console of the Central Manager can be remotely exploited by any attacker able to access the administrative UI via CVE 2024-21793 or CVE 2024-26026. This would result in full administrative control of the manager itself," the researchers explained. "Attackers can then take advantage of the other vulnerabilities to create new accounts on any BIG-IP Next asset managed by the Central Manager. Notably, these new malicious accounts would not be visible from the Central Manager itself."
F5’s NCM allows IT teams to manage devices such as application delivery controllers (ADCs), firewall solutions, and other network appliances. It provides capabilities for configuration management, policy enforcement, monitoring, and reporting across distributed environments. According to Shodan’s figures, there are more than 10,000 F5 BIG-IP devices with open management ports.
F5 also shared a workaround for admins who are unable to install the patch at this time. Per the company’s instructions, restricting Next Central Manager access to trusted users over a secure network sorts out the problem
There is no evidence of in-the-wild exploitation, Eclypsium confirmed.
The latest idea to cross our desks comes from Gabriel Ferraz, a computer engineer and TechPowerUp's SSD database maintainer, who turned a 512GB QLC SATA III SSD into a 120GB SLC one.
You probably know this, but just as refresher, SLC NAND holds one bit of data per cell, resulting in faster data writing, lower power consumption, and higher cell endurance than QLC NAND which stores four bits per cell. QLC NAND is denser and cheaper, but with the downside of compromised longevity and speed.
3000% endurance increase
Ferraz's idea was to trade capacity for massively improved performance and endurance. He took 512GB a Crucial BX500 SSD which has a Silicon Motion SM2259XT2 controller and NAND flash dies from Micron. Using an app called MPtools for the Silicon Motion SM2259XT2 controller, he identified the precise die used in the SSD and inputted in new die reference numbers.
Was it worth it? Well, while Ferraz lost a lot of drive space, he says “the SSD endurance jumps to 4000 TBW (write cycles), which is about a 3000% increase. Additionally, performance increased as well.”
Ferraz explains his process here, and you can also watch him perform his clever trick in the video below, which includes benchmarking results.
Nvidia has transformed into an AI superpower, becoming the third most valuable company in the world off the back of it, so it's perhaps no surprise other tech giants are looking on in envy and shifting their focus to follow suit.
During its recent earnings call, Samsung reported a consolidated operating profit of $4.8 billion in Q1 - a tenfold increase YoY - and company executives revealed a change in focus going forward.
The plan is now to concentrate on producing HBM and DDR5 memory and high-capacity SSD chips for the enterprise market, rather than targeting consumer PCs and mobile devices.
Meeting demand
"We plan to increase supply of HBM chips in 2024 by more than threefold versus last year," Kim Jae-june, Samsung’s memory business vice president, said on the call, reported the Korea Economic Daily. “We have already completed talks with our clients on this year's supply of HBM chips. In 2025, our HBM chip production will double from this year. Our talks on the 2025 volume with our customers are also going well.”
Samsung, currently ranked 23 in the world, has already invested heavily in HBM but currently trails behind its archrival, SK Hynix, in this area. SK Hynix recently announced plans to construct the world’s largest chip factory and has begun a partnership with Taiwanese foundry TSMC to produce HBM4.
Samsung said it anticipates a 50% increase in server DRAM production in Q2, and double output of server SSD in terms of bit growth. The company believes robust demand for AI chips will continue and stretch chip supplies throughout the year.
The South Korean tech giant also revealed it will commence mass production of its 8-layer HBM3E chips this month.
A prominent hardware leaker has alleged that while Thunderbolt 4 will come as standard for Intel Core Ultra 200 CPUs on Z890, that the upcoming CPU generation will miss out on Thunderbolt 5.
As a frame of reference, both Thunderbolt 3 and Thunderbolt 4 are capped at 40 Gbps which means there's no great increase in the two technologies despite their seven-year age gap. In contrast, Thunderbolt 5 can achieve double this at 80 Gbps which can be increased to 120 Gbps through Bandwidth Boost.
Also alleged by Golden Pig Upgrade is that Ultra Core 200 CPUs will feature just four Xe cores baked onto the chip which is half of what's currently available through Meteor Lake for laptops. This is unlikely to be too big a deal considering most (if not all) users will pair the processor with one of the best graphics cards, but it's worth noting.
We can take the alleged specs for the Arrow Lake flagship as a point of comparison. It's believed that the Intel Core Ultra 9 285K will feature 24 cores and 24 threads with a maximum boost clock of up to 5.5 GHz and a 125W TDP. In contrast, the current-generation Intel Core i9-14900K features 24 cores and 32 threads up to 6 GHz. It's certainly a different approach moving to Disaggregated architecture from Hybrid.
Not the best impression for Arrow Lake
Should Intel Arrow Lake miss out on Thunderbolt 5, that would mean that we would likely have to wait until Lunar Lake in laptops and Panther Lake in desktops to get ahead.
The new connectivity standard would mean not only increased bandwidth for external GPUs and SSDs but also boosting higher resolutions and framerates of up to 540Hz and enhanced multi-monitor in 4K and 8K (via Intel).
During Western Digital's recent Q3 earnings call, CEO David Goeckeler disclosed that the ever-growing need for higher capacity and speedier data access from customers across the world is pushing the company to expand its solid-state capacities.
The company chalked up a profitable quarter, with revenues soaring over forecast to $3.46 billion, a 29% YoY rise. The company managed to turn around a streak of losses, reporting a $135 million profit. These achievements are in stark contrast to rival Seagate, which posted an 11% YoY reduction in its revenues to $1.66 billion.
Goeckeler underlined that Western Digital's improved financial performance was a result of the company's efforts to offer a more diversified product range. He also said that WD was committed to delivering larger SSD capacities off the back of growing demand for AI-related applications. He said customers “want them [SSDs] in much bigger capacity points, 30- and 60-terabyte capacity points.”
HAMR HDD technology
Reporting on the third quarter results, Blocks & Files wrote “WD currently ships DC SN640 TLC PCIe gen 3 SSDs with up to 30.72 TB capacity and PCIe gen 4 SN650 and 655 drives with 15.36 TB. We now expect 60 TB SSDs to be announced by WD later this year.”
Without going into details of the exact capacities being worked on, Goeckeler said the company was expanding the size of the drives in line with what customers were demanding, stating WD is “increasing capacity and going through a qualification on that. So, we're in that process with customers.”
He also discussed hard-drive recording (HAMR) technology, including the issues surrounding it, stating, “we've been working on HAMR for quite some time. We understand HAMR extremely well. We understand all the issues with HAMR, and what it takes to get it qualified. Clearly, we're doing that all behind the scenes, because we have a product portfolio with the best TCO we can offer in the market today, and we can do that all the way up to 40 terabytes.” Western Digital’s rival Seagate recently announced the results of an experimental test that showed one of its hard drives using HAMR could run continuously for over 6,000 hours.
