TL;DR: Upgrade your PC to Windows 11 Pro for $14.97 before codes sell out (reg. $199). Fewer than 80 are left in stock. Microsoft has officially started the countdown clock on Windows 10. If you’re still procrastinating the upgrade, now is the time to act—before you’re left with no future security updates and paying full price for Windows 11 Pro. While codes last, you’re able to save a whopping $184. These Windows 11 Pro keys also upgrade two compatible PCs, like your personal and work computers, to the latest, most advanced operating system to date. Enjoy a refreshed user design, new productivity tools, improved gaming graphics with DirectX12 Ultimate, and enhanced security with Smart App Control and TPM 2.0. Since this is the Pro version of Windows 11, you’ll also get remote desktop access for viewing files and running programs from your mobile device. There is also BitLocker device encryption for protecting your hard drive in the case of device loss or theft, plus Azure AD, Hyper-V, and Windows Sandbox for professional use cases. Don’t miss this Windows 11 Pro discount—we don’t have many codes left at $14.97 (reg. $199). No coupon is needed to get this price. Microsoft Windows 11 ProSee Deal StackSocial prices subject to change. Read...

At a glanceExpert`s Rating Pros File, partition, and drive imaging to common VHD format Full-fledged partition manager Disaster recovery Optional retro command-line and character-based recovery interface Cons Needs to properly vet destination space Minor visual bugs and workflow peccadilloes Our Verdict R-Drive Image offers a ton of great backup features, including disaster recovery and newly expanded support for all the most popular online storage services. Price When Reviewed This value will show the geolocated pricing text for product undefined Best Pricing Today Retailer Price R-tools Technology $44.95 View Deal Price comparison from over 24,000 stores worldwide Product Price Price comparison from Backmarket Best Prices Today: R-Drive Image Retailer Price R-tools Technology $44.95 View Deal Price comparison from over 24,000 stores worldwide Product Price Price comparison from Backmarket I’ve truly enjoyed reviewing the various iterations of R-Drive Image — it’s been my go-to imaging/backup utility for nearly two decades. It’s a great program, and has still never created a bad backup in my hands. However, bugs, while rapidly fixed once reported, have reared their head in my last few looks. Including one that created an endless failure loop in 7.3. Come on now… What are R-Drive Image 7.3’s features? The big news for this dot upgrade is expanded support for cloud storage services — including the Amazon S3 and S3-compatible repositories, something I bemoaned the lack of in my review of version 7.2. Okay, bemoaned is a bit strong. Wished for? R-Drive Image’s main page. Also new in the cloud arena is support for Microsoft Azure and Google Cloud, though only in R-Drive’s pricier Technician and Corporate tiers. Most end-users will be more interested in Dropbox, OneDrive, and the cheaper S3 and S3-compatibles, which the consumer program supports flawlessly. The salient R-Drive Image features include: file, folder, partition, and whole drive imaging in both proprietary and non-proprietary VHD/VHDX formats; copying and cloning discs; a full-featured partition editor; and a bootable recovery disk with a choice between the modern and old-school, character-based interfaces. The R-Drive Image boot disk interface. The character-based interface (shown below) on the recovery disk is still my favorite for freaking out my friends and IT customers. Its appearance is as dated as my own, but it too is a hoot. Most users will prefer the modern interface shown in the previous image. You don’t have to use this retro-interface on the recovery disc. I do, but most users will use the one that mimics the Windows version. There are also Tasks, R-TT speak for jobs. These feature very granular scheduling and even media rotation in the pricier versions. Scripts are the same thing as tasks, but for command-line use. Yes, you can use R-Drive the old-fashioned way. Another feature I love in R-Drive Image 7.x is its replications. These are copies of the image saved to other destinations. In other words, you can define one data set to back up, and back it up to multiple locations using the same “script,” aka job. The R-Drive Image task schedular. If you’re looking for the ability to back up non-Windows computers, the recovery disc allows you to image FAT/exFAT, NTFS, ReFS / ReFS2+ (Resilient File System, Windows 2012/2016 Server), HFS/HFS+/APFS, XFS, Little and Big Endian variants of UFS1/UFS2 (FreeBSD/OpenBSD/NetBSD/Solaris), and Ext2/Ext3/Ext4 FS (Linux). Whew. That said, Apple, a tough-to-love company, makes it nearly impossible to boot from unsigned media these days, so backing up your Mac using R-Drive Image no longer works. I mention this because I tried and was disappointed that the FAT32 USB boot stick didn’t even show up. One R-Drive Image complaint is that while the program supports mounted network drives, WebDAV, and FTP, there’s no way to add files from a network location that’s not mounted as a drive letter by Windows. You can type in a network URL in the file name field for the destination though, and R-TT was asking me how to ameliorate this lack as I wrote this article. Expect this addition soon. My only other complaint is that the program will mildly warn you, but not stop you from trying to write an image to a destination that doesn’t have room for it. The latter became particularly apparent as a flaw during testing. More on that in the performance discussion below. You can define one data set to back up, and back it up to multiple locations using the same ‘script,’ aka job. How easy to use is R-Drive Image 7.3? For basic operations, R-Drive Image is quite easy to use. It walks you wizard-style through selecting files/folders, drives, or partitions, then where you want to write the resulting image. Creating recovery discs is simple as well. Where the program could use a bit of work is in the Tasks/Scripts area. The way they’re portrayed in the interface, it’s not immediately apparent that tasks are for the GUI and scripts are for command-line use. Yes, RTFM. You can specify media rotation options for any task. But only in the pricier professional versions of the program. I certainly appreciate that you can save ad-hoc jobs as command-line scripts, but it would be nice if you could save them as a task as well. Tasks, or recurring jobs, can only be created in the Tasks pane. This is also where you’ll find the retention options and scheduling. I’d also like to see an “Execute” button in the task bar. As it stands, you must right-click over the task to run it. R-Drive Image doesn’t make heavy use of right-click context menus, so this wasn’t something I immediately thought of. Double-clicking starts editing. I would’ve reversed those choices. How much is R-Drive image 7.3? The end-user version of R-Drive Image, good for one computer is $45. The transferrable-seat Technician version is $299, the Corporate version I tested is $189, and there are multi-seat and transferrable-seat versions for $500 and $900, respectively. You can see the full list and the feature differences at this link. How does R-Drive Image 7.3 perform? You may notice that this section of the review isn’t titled “How fast is…?” That’s because reliability and bug-free operation are every bit as important as how quickly backups are created. I’ve touted R-Drive Image as the most stable and reliable backup program I have ever used, and until version 7 — it was. It still is, when operations complete, but I discovered a couple of small visual glitches plus a rather large bug in testing 7.3 (7304) that I actually traced back to 7.0. The company fixed the issues within two days, which is fantastically responsive, but they should’ve been caught by the company before release. It’s not this error message that was disturbing, it was that clicking on cancel just brought it up again. Ending the program via Task Manager was required. Basically, before the fix, if R-Drive Image ran out of room on the destination media during a file imaging task it went into a closed loop with the error dialog that you couldn’t cancel out of. I had to end the program using Task Manager. Drive and partition imaging jobs showed the error dialog, but you could cancel out of it. Then there were the overlapping buttons shown below. Anyone using the program could spot them, which tells me that this version was likely rushed out the door. These overlapping buttons indicate that not enough people at the company are kicking the tires of new releases before they go public. Aside from the one fatal operational glitch and mis-drawn buttons, R-Drive Image performed as I’ve come to expect — admirably. Around a dozen operations to and from internal and external storage, NAS boxes, and online storage services (Dropbox, OneDrive, and Google Drive) went off without a hitch. The recovery media also booted just fine and backup/restore operations were glitch-free. This is more what I’m used to with R-Drive Image. In the final analysis, while the program had some issues with not enough space, all backups that succeeded retained their integrity and were mountable and restorable. That’s the most important aspect. Should you buy R-Drive image? Despite the glitches, which didn’t affect the ultimate sanctity of the backups, I still trust R-Drive Image’s backup engine. Additionally, with the new online support and multiple destinations per job, it’s one of the most complete backup solutions available at the price. I will continue to use and recommend it, but stop letting these faith-challenging bugs out the door, R-TT. Read...

Google has released Chrome version 134.0.6998.177/178 for Windows, while Chrome updates for macOS, Linux, and Android have yet to be announced. The security vulnerability closed by the update is apparently already being exploited by malicious actors. The makers of other Chromium-based browsers are likely to follow suit quickly. In the Chrome Releases blog post, Srinivas Sista names the eliminated vulnerability CVE-2025-2783, which was discovered by external security researchers and reported to Google. Google classifies the vulnerability as high risk. It’s an exploitable bug in Mojo on Windows that occurs under unspecified circumstances. (Mojo is a collection of runtime libraries for inter-process communication.) The vulnerability was reported to Google on March 20th by Boris Larin and Igor Kuznetsov, who are security experts at Kaspersky. According to Google, there are indications that this vulnerability is already being exploited for attacks in the wild. These indications are likely to have come from Larin and Kuznetsov, who presumably discovered the issue while analyzing malware attacks. It remains unclear who is attacking whom, but the attacks are apparently targeting Windows computers, as Google only provided this Chrome update for Windows. A week ago, Google released a security update for Chrome that closed two other security vulnerabilities. Chrome usually updates itself automatically when a new version is available. You can trigger the update check manually using the menu item Help > About Google Chrome. Google plans to release Chrome 135 in the coming week. The makers of other Chromium-based browsers are now required to follow suit with updates. Brave, Vivaldi, and Microsoft Edge should be updated with this security vulnerability fix later this week. Opera is still stuck on the outdated Chromium version 132. All of this once again highlights the importance of having good security software on your PC. Vulnerabilities exist everywhere and attackers can exploit them before app developers like Google are even aware of them. Antivirus suites offer an additional layer of protection, so check out our recommended PC antivirus software if you aren’t using one. Read...

The firewall is one of the most important security features in a computer environment. In the classic sense, a firewall is set up on an appliance, such as a computer that fulfills just a single, narrowly defined task — those found in companies, public authorities, and scientific institutions, for instance. A firewall of this type monitors the connection between two networks, such as the internal LAN and the internet, to prevent against unauthorized access. There are also personal firewalls such as the one that’s preconfigured in Windows, or available as freeware or as a paid software solution. In contrast to a classic firewall, a personal firewall does not protect an entire network, but only the individual computer on which it is installed. It’s the barrier that prevents any harmful elements from accessing the files that you have stored locally. The Windows firewall Microsoft calls the integrated Windows firewall the Windows Defender Firewall. This is the most important thing to know about the Defender Firewall: As the configuration of a firewall is a complex matter and requires a lot of prior knowledge from the user, Microsoft has defined three standard configurations for the Defender firewall. Which of these is activated depends on the selected network profile. If you select the public profile in the Windows Settings under “Network and Internet,” Windows seals off the computer from the outside world. In this case, the computer cannot be found on the network and it is not possible to share files and printers with other users. If you are not connected to a domain, you can choose between two network profiles in Windows, “Public” and “Private.” If you select “Public,” Windows largely closes the doors.Chris Hoffman / IDG However, if you set the private profile, you can share folders, files, and printers with others and use apps for internal communication. Windows also provides a domain profile for professional networks. Windows Firewall in practice The Windows firewall is active immediately after the operating system is installed and analyzes all incoming data packets to determine whether they have been requested by an application. This is called Stateful Packet Inspection (SPI), so the Windows firewall is a Stateful Packet Inspection firewall. However, programs such as Facebook Messenger or remote maintenance tools such as Teamviewer would not work if they were unable to receive unsolicited messages or calls. When they are installed behind a firewall, they therefore open one or more ports via which they receive incoming messages. They also customize other firewall settings. Readjustment by the user is not necessary. Teamviewer and other applications that react to signals from outside set the Windows firewall accordingly during installation. Foundry Verdict on the Defender firewall The Windows firewall protects the computer from malware that has infiltrated another PC and now wants to spread via the network. It is therefore strongly recommended not to deactivate it. However, at least in the default setting, it cannot prevent malware that has managed to infect one computer from trying to spread it to other PCs. Specialized firewall applications can do this better. Firewalls from third-party manufacturers The Comodo firewall can block applications, hide the PC in the network, and list which programs are currently connected to the internet. However, it only comes in a complete package with an antivirus program.Comodo If you travel a lot and use your notebook to connect to other people’s LAN and WLAN networks, you naturally have no control over the router settings. In this case, we recommend using a third-party firewall. The Comodo Firewall, which is sold as part of the Comodo Internet Security package, is widely used. Unfortunately, the package can only be installed as a whole, and the setup also includes the Comodo antivirus program. Once the installation is complete, click on the shield icon labelled “Protection” on the left-hand side and click on “Firewall” in the next window. You can then prohibit or allow applications to connect to the internet, you can block incoming connections via port blocks or have them reported, you can manage your network connections and the active connections to the internet. If you are travelling and in a foreign WLAN, for example, it is advisable to hide the ports and only allow the applications that you actually need. In some cases, you may need to allow several program files to access the internet. You can often find information on this from the manufacturer or in the help forums on the internet. Mobile PCs require a more comprehensive protection function than stationary computers. A more configurable and feature-rich firewall from a third-party manufacturer is recommended for them. Sandbox for isolation Sandboxes are a component of many personal firewalls. Under “Protection,” you can access Comodo’s sandbox function, which is called “Containment” here. A sandbox is used to run individual, selected applications in a closed memory area where it cannot cause any damage to the rest of the PC. In a sandbox, for example, suspicious downloaded EXE files and other programs can be executed without the user having to worry about catching a virus. You must first activate the Windows sandbox via “Programs and Features” in the Control Panel.Foundry Windows Pro also has a sandbox, but you have to activate it first. (It’s not available for Windows Home.) To do this, open the Control Panel and go to “Programs and Features” in the icon view. In the following window, click on “Enable or disable Windows features,” scroll down, and tick the box next to “Windows sandbox.” get windows 11 pro for cheap Windows 11 Pro After confirming with “OK,” Windows installs some files and then reboots. After logging in, you will see a window with a second Windows desktop. To test programs there, you can copy them to the sandbox using the clipboard. You can close the sandbox window like any other window. To open it again, enter Windows Sandbox in the search field of the taskbar and press Enter. Further reading: Be safe! How I set up a ‘Paranoid PC’ with Windows Sandbox to surf the risky web Professional firewall for home use Professional firewalls usually run on appliances, i.e. their own hardware, and therefore cost several thousand dollars. However, there is a cheaper option: The free open-source firewall Ipfire is also available for the Raspberry Pi from version 4b with 1GB RAM and more. The software can shield entire networks from the outside world, is easy to use, and even offers an optional intrusion detection system (IDS) that monitors for changes that are typically caused by malware. With IDS, however, the memory expansion should be 4GB or more. Read...

At HP Amplify 2025, the company’s big annual conference for showcasing its latest products and services, HP unveiled nearly the entire set of its new PCs for the year. I lost count at some point, but HP claims over 60 new models of laptops and PCs. While technically true, it’s a bit fudged—the company counts some variants of the same computer as separate. For example, if the same laptop comes in Intel, AMD, and Snapdragon options, then each one is a distinct “model” even if everything else is the same. That also goes for screen sizes, 2-in-1 variants, and so on. The calm before the storm, just before HP Amplify 2025 kicked off.Mattias Inghe / Foundry Even so, there were a lot of computers. I was there in person, wandering around the showroom with everything on display, along with plenty of other innovations (especially for business customers), such as management systems, security and printing solutions, etc. But I was there for the computers, from neat little desktop PCs to mighty workstations to laptops of all kinds. There were so, so many laptops. OmniBook takes over the laptops It was hard to tell them apart, but HP aims to make them easier to distinguish than before. The older, confusing sub-brands—Pavilion, Envy, and Spectre—have been scrapped and replaced by OmniBook. HP ushered in that line with a few OmniBook laptops last year, but now everything will be OmniBook going forward. These OmniBooks are then divided by suffixes. For example, the OmniBook 3 is cheap and simple, the OmniBook 5 more “normal” and mid-range, the OmniBook 7 more luxurious and lighter, and the OmniBook X supposedly at the top… except for the OmniBook Ultra, which is really the high-performance tier. Lots of new business laptops being shown off at HP Amplify 2025.Mattias Inghe / Foundry Are HP’s laptops now easier to parse? Maybe. Maybe not. I’ll have to come back in a year or two when all the older models have disappeared and OmniBook reigns alone. But to make matters a bit more confusing, there’s also the new OmniDesk line of desktop PCs and OmniStudio line of all-in-one machines. HP’s gaming PCs retain the names Omen and Victus (with only a single model launched in the latter). Business laptops are still called ProBook and EliteBook, but specialized model names are apparently being phased out here as well. (I didn’t see any new “Dragonfly” ultra-light business laptops, for example, but some EliteBooks were really slim.) I saw all kinds of consumer PCs, like these stylish desktop OmniDesk models.Mattias Inghe / Foundry Finally, there’s the Z series of powerful workstations for professionals, which get to keep their names. Plenty of new ZBook laptops and Z desktop PCs were also on show. AI is creeping into everything It’s no secret that HP is making a major investment in AI. Expect all new releases, starting with the OmniBook 5 and up, to be equipped with an NPU that can run local AI features. What about the budget-focused OmniBook 3? It’s unclear whether that one will also come with AI capabilities. I didn’t see any, but maybe with time. And even more computers! It just never ended.Mattias Inghe / Foundry I couldn’t quite figure out if there was Copilot+ certification across the board, but HP has added several of its own AI features to its machines, right down to the cheapest models. For example, there are Poly Studio (which adds webcam effects and controls, plus noise reduction for microphones) and AI Companion (HP’s own AI app that indexes local documents and lets you do semantic searches via a chat interface). All of this happens locally on your computer’s NPU to keep your data private and secure. My 5 favorites from all the ones I saw Mattias Inghe / Foundry HP Omen 16 Slim. The new gaming laptop that’s barely 0.89 inches thick and only weighs 5.29 pounds but has a fast 16-inch display, massive cooling elements, and heavy performance under the hood with an Intel Core Ultra 9 285H processor and Nvidia GeForce RTX 5070 graphics. Learn more about it in PCWorld’s hands-on experience of it. Mattias Inghe / Foundry HP OmniBook 7 Aero. A very stylish 13-inch laptop that weighs about 2 pounds and packs plenty of performance thanks to its AMD Ryzen AI 300 processor. It has decent graphics performance, 50 TOPS for AI, and low-power operation. I’m really looking forward to trying this one. Mattias Inghe / Foundry HP OmniBook 5. This one comes in sizes from 14 to 16 inches and feels very stable, comfortable, and well-built for something that starts around $800. That’s a great price for a quality laptop with AI performance, and it could be many people’s first AI PC. Mattias Inghe / Foundry HP Z2 G1a. When I tested Asus’ monster ROG Flow Z13 tablet, I wondered what a compact desktop PC with the same processor (Ryzen AI Max Pro 395) could do with more cooling and higher power. The answer is here. HP has put it in a compact mini PC format and is selling it as an HP Z workstation. It really does go fast—and it costs. The AMD rep I spoke to said it would be released globally, but couldn’t say for sure if it was for all markets. I hope so because I want to test it. Mattias Inghe / Foundry HP EliteStudio 8. This all-in-one computer seriously lives up to its “all in one” epithet. With a large, built-in Poly webcam that pops up, speakers built for clear voices, and AI-powered microphone management, along with Poly Studio software, this machine is perfect as a video conferencing workstation. With its last-gen Intel Core Ultra processor, AI performance isn’t quite up to scratch, but it’s good enough for most tasks. KVM functionality means you can also use it to dock your laptop. It certainly seems complete and very all-in-one. There could of course be more highlights in the deluge of computer news. If I were to ask HP, they’d say every single one was a highlight, so I was none the wiser. But these made my tester fingers itch and I hope all of them can be found and reviewed during the year. Read...

Far be it from me to tell the most valuable company on the planet what to do. But Nvidia, you do realize that your whole schtick is graphics cards, right? You know, GeForce, those chunky go-fast gadgets you’ve been selling for 25 years or so, to let people play video games? Because based on some recent news, they don’t seem to be the company’s biggest focus. New Nvidia-designed Arm desktops After a couple of months of graphics card launches that have been frustrating at best and a complete retail disaster at worst, Nvidia CEO Jensen Huang took to the stage at the company’s self-branded GPU Technology Conference. Now, to set the proper tone for a bit of admittedly entitled whinging, Nvidia has rebranded the initial portion of this yearly event as an “AI conference,” so it’s hardly surprising that AI computing is the lion’s share of the two-hour keynote talk. Even so, I can’t help but feel a little miffed that Huang fawned over new Nvidia-branded desktop computers it’ll be selling directly to the AI industry and individuals. The DGX Spark mini-PC and DGX Station desktop are based on the new Grace Blackwell AI-focused GPUs. This sort of setup was previously available only in datacenter hardware. The specs for the larger DGX Station aren’t available yet, but the Blackwell GPU in the smaller Spark (previously revealed as Project DIGITS) will be paired to a custom Arm CPU. Nvidia Despite being loaded with up to 128GB of memory and 4TB of storage, this thing isn’t going to make a good fit as a conventional desktop — and at a $3000 pre-order price, it wouldn’t find many buyers anyway. Nvidia is positioning both the DGX Spark and DGX Station as all-in-one solutions for running local AI models, along with a token mention of “researchers, data scientists, robotics developers, and students.” Nvidia will be partnering with Asus, Boxx, Dell, HP, Lambda, and Supermicro to manufacture and sell both desktops “later this year.” Didn’t Nvidia use to sell graphics cards? Again, it’s not as if it’s illegal or even distasteful to for Nvidia to branch out into other markets beyond gaming chips for PCs, or even beyond gaming entirely (remember that Nvidia powers the Nintendo Switch with Tegra mobile chips, after abandoning the smartphone market over a decade ago). And Nvidia hardware is in a lot of things you might not expect, including connected cars, medical devices, and other industrial hardware. But as a PC gamer myself and an advocate for other gamers, I can’t help but feel a little betrayed. While retailers, AIB partners, and scalpers are eating good on the scarcity of new GeForce cards, Nvidia is preparing entirely new product lines for the AI industry. And not just the “big iron” hardware that’s made it an unstoppable juggernaut of tech over the last few years, relatively small, targetted devices for individual users. Adam Patrick Murray / Foundry Nvidia can make both, of course…but even a big green giant only has so much chip production capacity. I don’t have any empirical evidence to prove it, but I’m not the only one who feels like the woefully inadequate supply of new RTX 50-series graphics cards has something to do with all these new industrial chips and brand new Arm processors being made for customers with even deeper pockets. It doesn’t help that Nvidia has been evasive, if not entirely deceptive, in addressing its supply issues as RTX cards vanish from shelves the second they arrive. Nvidia jumped from one peak to the next by riding the crypto boom right into the AI boom. And while recent developments have caused it to look more like a bubble, Nvidia stock is still more than double what it was at the beginning of 2024. The cold calculus of profit suggests that the company is far more concerned with AI, for the present if not the future, than it is with PC gaming. As my grandpa used to say, make hay while the sun shines. Opportunities for AMD and Intel But that focus can’t help but alienate PC gamers, and it sure doesn’t help that Nvidia seems to be making a lot of money off of that, too. It creates an opportunity for the competition. If AMD can’t make infinite money off the AI industry, it can sure sell some graphics cards, especially to the large chunk of PC gamers who can’t afford four figures for a new GPU. While Nvidia still dominates to the point of monopoly for discrete graphics, recent sales data indicates that AMD is clawing back some market share, especially in markets like Japan and Taiwan. And if AMD is making a play for the mid-range, newcomer Intel is doing the same thing for the budget space with its second generation of Arc cards. Adam Patrick Murray / Foundry Nvidia announcing new AI PCs isn’t necessarily at the expense of its ability to deliver gaming GPUs. But it isn’t happening in a vacuum, either…and Nvidia is failing to deliver gaming GPUs. If you’re considering an upgrade or a new purchase, maybe give the competition a try. It’s not as if Nvidia even needs a desktop GPU monopoly anymore. And if the AI bubble bursts, and the company is no longer flush with cash…a little ego bruising might be good for it when it decides that it wants to sell graphics cards to PC gamers after all. Read...

The time has come! Microsoft is now sending out emails to Windows 10 users reminding them that end of support for the still-very-popular operating system is imminent. Users have begun receiving the first of these emails in their inboxes, reports Windows Latest. The emails begin with a heading that boldly proclaims “End of support for Windows 10 is approaching.” Below that, there’s a button to “check upgrade eligibility” (presumably for bumping your PC up to Windows 11) as well as a link to “Explore new computers” (for users who are still on older PCs that aren’t compatible with Windows 11). And below that, ther’s an FAQ on what Windows 10 end of support means. Get Windows 11 Pro for cheap Windows 11 Pro While Microsoft explains that Windows 10 will no longer be supported after October 14, 2025, the FAQ fails to mention that Microsoft is offering paid extended support for at least another year. The FAQ also gives some advice on what to do with a retired Windows 10 PC, emphasizing that Windows 10 PCs won’t simply stop working after end of support hits but will grow increasingly more susceptible to malware and hacker attacks because the company won’t be providing security updates that address newly discovered security flaws. The FAQ ends with the inevitable claim that Windows 11 is the most secure Windows ever and that you should please upgrade to it. How to continue using Windows 10 As mentioned above, Microsoft’s email fails to bring up the company’s offer of one year of extended support for a $30 fee. Alternatively, you can use 0Patch to keep getting updates after support ends. If neither suits your fancy, you can also look into some other ways to save your PC once Windows 10 hits end of life. What you shouldn’t do is keep using Windows 10 without further security updates as that’ll leave you exposed and vulnerable to all kinds of cybersecurity shenanigans. If you decide to do this anyway, make sure you’re protected by reputable antivirus software and VPN services. And if you continue using your Windows 10 computer, Microsoft also advises that you back up your files to OneDrive. Read...

I could spoil this entire article and tell you right off the bat which high-speed bus protocol you should be looking at for you current and future external storage needs… But where’s the fun in that? If you need an answer right away, simply scroll to the bottom and grab your instant gratification. Otherwise, educate yourself with this quick treatise on the various high-speed external protocols that are available, as well as their respective strengths and weaknesses. 10Gbps USB (USB 3.2) So, have we really come so far that a 10Gbps/1GBps transfer rate seems quaint? I’d say, yes, at least from the perspective of a long-time storage reviewer. Nevertheless, for small amounts of data, a 1GBps transfers rate is still going to get the job done quickly enough. SK Hynix 10Gbps Tube T31 SSDJon L. Jacobi And, of course, for a lot less cash than the 20Gbps/40Gbps options. If capacity means more to you than top performance, then 10Gbps USB (or even 5Gbps if you really don’t care about speed) could be the way to go. Best Usb 3.2 external drive Crucial X9 Pro Read our review Best Prices Today: $95 at Amazon 20Gbps USB (USB 3.2×2) This protocol offers fast 20Gbps transfers, but it’s the odd duck of the bunch. It will work on normal USB ports at reduced 10Gbps/5Gbps speed, but a dedicated and not particularly common 3.2×2 port (sometimes referenced at 20Gbps) is required to attain the promised 2GBps transfers. Another downside is that unlike USB4, USB 3.2×2 will not function at its full 20Gbps on most Thunderbolt 3 or 4 ports, instead dropping to 10Gbps — the rate of most native USB implementations. Thunderbolt passes through USB to the system. That speed drop afflicts all new Apple Silicon Macs. I was sorely disappointed when I found out that my more-than-fast-enough 20Gbps USB SSDs transferred at half speed on my otherwise stellar-performing Mac Studio. Crucial’s 20Gbps X10 ProJon L. Jacobi On the plus side, 20Gbps USB SSDs run cooler, use less juice than their 40Gbps counterparts, and still gets things transferred pretty darn quickly. The price of USB 3.2×2 SSDs has dropped some recently, due to USB4 showing up, but they’re still not as affordable as a 10Gbps SSDs. best usb 3.2×2 External drive Lexar SL600 20Gbps USB SSD Read our review Best Prices Today: $129.99 at B & H Photo | $150.24 at Amazon Thunderbolt While Thunderbolt has been the de facto high-speed port on Macs for well over a decade now, it’s not nearly as widespread on Windows PC platforms. Largely that’s because it’s pricier to implement than USB thanks to royalties and greater power delivery. That goes for both motherboards and drives. Thunderbolt is certainly more common than it once was, but it’s hardly ubiquitous on the Windows side of the pond. Here are the various flavors you might run across. Thunderbolt 1/2: These older versions with Mini-DisplayPort connectors and 10Gbps/20Gbps transfer rates are only found on legacy equipment, though you can connect newer Type-C Thunderbolt peripherals to these ports and vice versa using Apple’s Thunderbolt adapter. There are others, but check the specs carefully — some don’t connect 1/2 to 3/4/5, only vice versa. My advice? Stick with newer Thunderbolt 3/4/5 peripherals and use the adapter. Or perhaps upgrade your computer. Glyph Atom Pro SSD Jon L. Jacobi Thunderbolt 3: This version of the technology doubled bandwidth to 40Gbps and was the first version to use the Type-C connector. The latter confused just about everyone, as USB switched to this connector around the same time. Thankfully, Thunderbolt 3 tunnels USB so you may attach USB peripherals to a Thunderbolt port, if not the other way around. Thunderbolt 3 can achieve around 3GBps during transfers on most computers sporting the technology. It’s also getting cheaper as the industry moves on to Thunderbolt 4/5. Thunderbolt 4: 40Gbps Thunderbolt 4 was more a certification update than a change in the technology. Intel found that vendors weren’t implementing all of Thunderbolt 3’s features all the time. Thunderbolt 4 made implementing the entire stadard mandatory if you wanted to use the name and logo. Thunderbolt 5: Version 5 doubles version 3/4’s 40Gbps storage bandwidth to 80Gbps. You may see 120Gbps mentioned in slides, but that’s downstream only to drive displays. Thunderbolt 5 also delivers significantly more power across the bus at 240W as opposed to Thunderbolt 3/4’s 100 watts and the mere 9.9 watts of 1/2. This makes Thunderbolt 5 the go-to for power users with power hungry storage peripherals — e.g., a RAID box. OWC’s 80Gbps Envoy Ultra Thunderbolt 5Jon L. Jacobi But wait… There’s a caveat! While Thunderbolt 5’s specs sound spectacular and it benchmarks great, it hasn’t shown a significant impact on real-world performance in our real-world large file transfers. Also, the doubled bandwidth does nothing to speed random performance, which is far more about seek time than raw throughput. So it won’t run an operating system appreciably faster than an older Thunderbolt 3/4 SSD. USB4 As you may or may not be aware, USB4 is basically a version of Thunderbolt 4 that was donated to the USB Forum by Intel. Hence there’s an awful lot in common including the up-to-40Gbps transfers. The ace up USB4’s sleeve is far greater compatibility. While you can’t attach Thunderbolt peripherals to older 1/2/3.x USB ports, you can attach USB4 peripherals to those ports with every expectation of them operating. Of course, only at the maximum speed of the older port — but hey, better slower than not at all! Also, the shared heritage means you can attach a USB4 peripheral to a Thunderbolt 3/4/5 port and have it work at its best speed: 20Gbps on Thunderbolt 3 and 40Gbps on the other two. Adata’s 40Gbps SE920 USB4 SSDJon L. Jacobi If almost complete compatibility weren’t enough, USB4 seems to be coming to market at a lower price point than Thunderbolt 3/4/5. Adata’s uber-fast SE920 is roughly $125 per terabyte, about par with no-name Thunderbolt 4/5 SSDs. Alternatively, you can grab a USB4 enclosure such as UGreen’s CM642 for less than $100 and add your own NVMe SSD. Note that there is an 80Gbps USB4 2.0 in the works (of course there is!), but we haven’t seen an implementation yet. best usb4 external drive Adata SE920 USB4 SSD Best Prices Today: $179.99 at Amazon USB4 for the win! If you just read all of the above, it shouldn’t surprise you that I’m giving the nod to USB4. It’s very fast, compatible with just about every computer and device on the planet, and more affordable than Thunderbolt. The only exceptions to that recommendation would be 5/10Gbps USB if you need something super cheap, or Thunderbolt 5 if you want absolute top performance and have the software and hardware to take advantage of it. Further reading: Best external drives Read...

If you need to “remote” in to a local PC while on vacation. there’s a new way to do it: the Windows App, which is replacing the Remote Desktop app within Windows. Microsoft’s Remote Desktop app is one of the easier ways to remotely connect to a PC while you’re out and about, allowing you to access a PC remotely and even perform some basic file transfers. (Unfortunately, since the October 2024 Update for Windows 11, the app has had some issues.) In September 2024, however, Microsoft began rounding up some of its connectivity services and putting them inside what it calls the Windows App. (Yes, we think that the name isn’t the best, either.) However, Microsoft did so as a unified way to allow users to connect remotely from other PCs or mobile devices. Now, having migrated users over to the Windows App, Microsoft is removing support for the Remote Desktop app, it said this week. “Starting May 27, 2025, the Remote Desktop app for Windows from the Microsoft Store will no longer be supported or available for download and installation,” Microsoft said. “Users must transition to Windows App to ensure continued access to Windows 365, Azure Virtual Desktop, and Microsoft Dev Box.” The process has some nuances. If you’re a Remote Desktop user, you’ll need to keep using the Remote Desktop Connection until this feature is in the Windows App, Microsoft said. If you use Remote Desktop Services, you should use the RemoteApp and Desktop Connection until support for this connection type is available in Windows App, the company said. Everyday PC users who need to remote into far-away computers might be better off with one of our favorite remote desktop services. Read...

An Australian company called Cortical Labs has developed a computer powered by lab-grown human brain cells, Gizmodo reports. The computer, known as CL1, is described as the world’s first “code deployable biological computer” and is now available for pre-order — for a price in the $35,000 range. Don’t want to buy your own device? The company also offers “Wetware-as-a-Service” via which you can rent bio-computer processing power via the cloud. CL1 consists of lab-grown neurons grown on a glass-and-metal electrode array. They’re connected to 59 electrodes, creating a stable neural network. The system is encased in a life support unit that keeps the neurons alive by mimicking the body’s organ functions, including heart pumping, kidney-like waste filtration, and gas mixing of oxygen, carbon dioxide, and nitrogen. According to Cortical Labs, the neurons are placed in a nutrient solution and receive their information from the company’s Biological Intelligence Operating System (biOS), which creates a simulated world in which the neurons receive sensory input and produce responses that affect the environment. CL1 is designed as a high-performance closed loop, where neurons interact with software in real time. The system can stay alive for up to six months and is compatible with USB devices. Cortical Labs demonstrated an early version of the technology by teaching the system to play Pong. They claim that biological computers can rival or surpass digital AI systems, especially when it comes to understanding the basic mechanisms of intelligence. According to the company’s Chief Scientific Officer, Brett Kagan, a network of 120 CL1 devices could give researchers insight into how genes and proteins affect learning. The technology can also be used in drug development and disease modeling by simulating neurological processes at the molecular level. Read...