According to the latest Valve report (participation in the survey is optional), as of early January 2026, 70.83% of all survey participants run Windows 11 on their systems. In only one month, the operating system jumped by 5.24 points, reaching a new record high. As for Windows 10, it dropped by 2.36 points to 26.70%. Although Windows 10's share is reducing, the operating system will most likely stick around for a while, given that many users cannot update their PCs officially (you can try this guide) and building a new PC right now is quite a task in light of astronomical memory and storage prices.

December 2025 is also the first month to lack Windows 7 in its stats. Now, the Windows part in the OS section has only Windows 11 and 64-bit Windows 10, which, when combined, total 94.95% of all survey participants (+0.16 points).

Linux is the second most popular operating system on Steam. Calling it popular is a bit of a stretch, given it only has 3.19% (-0.01 point), but the ever-growing dissatisfaction with what Microsoft is doing to Windows may change things in the long term. We have already published a more detailed breakdown of Linux performance on Steam, and you can check it out here.

Finally, macOS is third with a modest 1.86 (-0.16 points), with macOS 26 Tahoe being the most popular release.

When it comes to hardware, the most popular picks include 16GB of memory (40.14%, -0.8 points), a six-core processor (30.8%, +2.36 points), the NVIDIA RTX 3060 (6.26%, +2.1 points), 8GB of video memory (32.53%, -0.83 points), and a FullHD monitor (53.68%, +0.85 points). You can find more information on the official Steam website.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Saturday 3 January 2026 at 4:34 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Do you remember when using a computer, especially Windows, felt fun? It might be just me in "man yells at cloud" mode or overdosing on Windows 7 nostalgia (I will rant quite a bit), but I remember how exciting it felt to go through the initial setup in Windows Vista, 7, heck, even Windows 8/8.1. These philosophical ponderings gave an idea to compare the initial setup experience of all the operating systems that Microsoft released over the last two decades, and see if things that had been added over the years are actually useful stuff or just bloat and ads (they are).

Windows Vista (2006)

The first thing you see after installing Windows Vista is a prompt to create a profile, set a password, and select a profile image. Things are simple, easy to understand, and the built-in pictures are diverse and cute. Somehow, the default flower was always my favorite.

Next, we get another piece of personalization, as you are asked to give your PC a name and select a desktop background. Again, Vista sets priorities straight and encourages the user to do fun things like choosing a profile picture or the desktop background. Only then do we proceed to Windows Update settings with clean and easy-to-understand options. Do not want Windows updates? No worries, man, but keep in mind security risks.

The final two steps involve setting up the time/date and your network type. After that, Windows Vista presents you with its beautiful logo and polite "thank you." Overall, 5 pages, zero ads, and no user-hostile stuff or dark patterns. 10/10 OOBE experience.

 

 

 

 

 

Windows 7 (2009)

Windows 7's initial setup is mostly the same, but a little more serious. You no longer get to choose your default background or profile picture. Just give your PC a name, set a profile with a password, select your network type, select the right Windows Update setting, and adjust the date. Quick and easy, still good, but not-so-fun anymore.

 

 

 

 

 

Windows 8 / 8.1 (2012)

Things change a lot in Windows 8/8.1. Personalization is back, as you get to choose a theme for the entire experience and rename your PC from the get-go. Microsoft is now trying to make things easy for you by suggesting express settings, but we are taking the longer route here, so let's click Customize.

Windows asks if you want to make your PC discoverable for devices within your network, and then you choose if you want to auto-update your apps and get drivers for new devices. There are also two SmartScreen security options, which are turned on by default, but you can disable everything if you wish.

On the next step, Windows offers more useful options like finding solutions for errors or optimizing the web experience in Internet Explorer. These are turned on by default (I do not mind that), but all the privacy features for "improving Microsoft products" are disabled by default, which is nice.

However, on the next step, more stuff that sends data to Microsoft is enabled by default, so you'd better read what those toggles mean.

Finally, Microsoft asks you to sign in with a Microsoft Account. There is no apparent "Offline account" button, which is not cool, but you can click "Create a new account" and then proceed with an offline one, something that in Windows 11 Microsoft fears more than a plague.

In total, we have 8 pages of settings, with mostly harmless stuff and a bunch of features that you may want to turn off. Overall, good experience.

 

 

 

 

 

 

 

Windows 10 (2015)

I am looking at the initial version of Windows 10, version 1507. To my surprise, the entire setup process is simpler and more compact than Windows 8's. However, everything is now enabled by default, and there are more eyebrow-raising features privacy-wise. As for using a Microsoft Account, we now have a clear "Skip this step" button, which I, for one, see as a welcome addition that Microsoft somehow managed to keep mostly intact for a decade of Windows 10's lifespan (it later evolved into "Offline account"). Impressive.

In total, seven pages, zero ads.

 

 

 

 

 

 

Windows 11 (2025)

Now, let's talk Windows 11, and oh boy, things get messy here. We move from a quick and simple initial setup to a process that desperately needs an ad block. Everything begins mostly alright. You select your language and region settings, give your PC a name, set up a pin, glance over privacy settings, all of which are toggled on, etc. Among those steps, Windows 11 starts checking for updates without asking you, so the process is taking notably more time.

After asking to sign in with a Microsoft account without alternatives (Microsoft is actively fighting against users who want to have local accounts), Windows 11 presents you with an option to restore your PC from a backup without an obvious "no" button. The only way to set it up as a new PC is to click "More options" > "Set up as a new PC." Both options conveniently look like links, not buttons, which are the first signs of dark patterns to persuade users not to use them.

When you click "Set up as a new PC," Windows 11 still nags you to restore from a backup, just blatantly disrespecting your choice.

The next screen asks you to customize your experience. This thing always baffled me, as it is just a way to tell Microsoft what apps to pin to the taskbar right from the get-go, i.e., show ads. Fortunately, there is a Skip button.

Next, Windows 11 proposes that you connect your Android phone to your PC using Phone Link. It is a useful thing, so I consider this a useful suggestion. What follows next is a OneDrive ad for your phone, which feels a bit random, like, why do I need to set up photo backup on my phone when I set up my PC?

Next, Microsoft tries to hijack your browser data in advance by asking you to allow Windows to pull data from third-party browsers and inject it into Edge under the pretense of helpful features and "saving time and money" (that thing actually steals affiliate links from creators, by the way). Ugh. Not now, Microsoft.

The following three screens nag you with Microsoft 365 ads. First, Microsoft tries to lure you with a "you are in luck" message to sign up for a free trial of Microsoft 365. We are declining this, but they do not give up and try to harvest at least $1.99 with a Microsoft 365 Basic proposal. Again, decline. Now we get another proposal, this time to use Microsoft 365 for free with the universally hated new Outlook client for Windows (conveniently turned on by default).

Finally, one more ad: the ever-growing in price Game Pass for only $14.99. Skip, and you are on the desktop at long last.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Overall, 18 screens with a lot of "getting things ready for you" spinners, ads, and awkward button placement. Out of those 18 screens, 7 were just ads. And while you can skip them, Microsoft won't give up on soliciting, with Game Pass, Edge, and Microsoft 365 ads sprinkled all over the operating system. Where is customization, where is the ability to select light or dark mode, where is the option to change the desktop background, where is the fun, Microsoft?

Yes, I know that tech evolves, and Windows 11 is now much more capable than its predecessor. Some of that stuff is useful. Windows Backup is alright, as long as it does not treat me like an idiot by hiding the decline button behind two additional screens. Phone Link is also useful, but that is all. Everything else is just ads, as Microsoft desperately tries to boost its KPI and squeeze every penny from its users. Ah, yes, do not forget that your drives are now encrypted without your consent, so better back up that BitLocker key now.

Setting up Windows nowadays feels like fighting off an annoying salesman instead of quickly setting up basic stuff before getting to the desktop, not to mention all the user-hostile practices like not allowing you to proceed with the setup if you do not have the internet and do not know the magic command, or not allowing you to use a local account. And in the world of Windows updates getting crappier with each month, the lack of Windows update controls makes it even worse. I know that Windows updates are important in the modern world, but at least give us basic controls like not installing updates or drivers right here and right now. Old Windows versions respected your choice, but now, we get three screens begging you to use Windows Backup.

With an approach like this, it's no wonder people grew tired of Microsoft, and the remaining fans struggle to keep the dying flame of enthusiasm. Sad times.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Friday 2 January 2026 at 4:30 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

While I share the pessimistic sentiment about Windows ens*ittification, in 2025, Windows 11 received quite a lot of new and useful stuff. In this article, I want to highlight my top 10 changes and improvements. Some of these are quite big, some are minor, but they surely made the operating system better than before.

Note: Features from this list may still be rolling out gradually.

1. Start menu

This is undoubtedly the biggest and most important change for Windows 11, as the Start menu has been among the biggest complaints since day one. Microsoft finally brought more layouts with more efficient screen use, the long-requested ability to disable recommendations properly, and more.

The new Start menu is not perfect, though. You still cannot resize it, and there are some odd bugs, but for what it is worth, it is a massive upgrade over the original, poorly implemented menu from 2021. I personally do not use the Start menu at all (Command Palette single-handedly killed Start for me), but I understand how integral it is for millions of users, so I appreciate Microsoft doing the right thing.

2. Redesigned battery indicators

They are awesome! The battery indicator has been annoying me since day one. It was too small, too hard to read, and lacked the ability to display battery percentage. After years of complaints from users, Microsoft finally redesigned the battery indicator on the taskbar and the lock screen. The new one is bigger, has colorful indicators for charging and low-power mode, and lets you display battery charge in percentages.

3. Taskbar upgrades

The new battery indicator is not the only taskbar upgrade in Windows 11 this year. Microsoft finally fixed the stupid inability to open the notification center on secondary displays. Small icons are also here, and you can have a dedicated button for the emoji panel, which is a handy alternative to the standard shortcut (Win + ;).

Again, not everything is perfect. You cannot have a smaller taskbar when using small taskbar icons like in Windows 10, and Microsoft still won't let us move the taskbar to the top or sides of the screen. Come on, Microsoft!

4. Better dark mode

At long last, Microsoft equipped common file dialogs with dark mode support so that the system no longer blinds you whenever you delete a file, copy/move something, etc. Microsoft is finally taking dark mode in Windows 11 more seriously, and I am very happy about it. If only Windows 11 knew how to switch modes on schedule, but this omission is well taken care of by the new Light Switch module in PowerToys.

5. Em dash shortcut

This one is small but very handy. Now I can press Win and minus to enter an em dash. As a person who uses the em dash often, this tiny change improved my typing experience a lot.

6. Windows Backup App

In 2025, Microsoft made the Windows Backup app a lot more useful; now, it can transfer your files and settings from your old PC to a new one, making the transition a lot more seamless for those who have a bunch of local files and no means to transfer them to a new computer.

7. Gamepad keyboard

Windows 11 now has a great on-screen keyboard for your gamepad, which makes it much easier to type something when gaming without reaching for your physical keyboard. It is a great implementation, with keys properly mapped to controller buttons so that you can quickly type whatever you need from your coach. It is a great addition to the gaming experience alongside stuff like the built-in browser in Game Bar.

8. Accessibility

Microsoft pays quite a lot of attention to accessibility in its operating system, ensuring everyone can use Windows 11 regardless of their physical abilities. In 2025, Windows 11 received upgrades to Voice Access and typing, Narrator, and Magnifier.

Windows 11 now uses AI to clean up your speech and correct mistakes on the go, it supports conversational commands, and works with custom vocabularies. You can also adjust wait time, disable profanity filter, and more.

Narrator can now generate richer image descriptions, read more smoothly in Word, protect your privacy with Screen Curtain, and utilize HD voices that sound more natural and expressive. Finally, Magnifier now features faster, clearer navigation with one-click zoom controls and HD voices for a more natural and engaging experience.

9. Quick Machine Recovery

Quick Machine Recovery is a new tool that allows Microsoft to deploy fixes for outages quickly and efficiently. If another botched update takes down a chunk of Windows systems and they can no longer boot, Microsoft can send an emergency update and bring those PCs back to life from the Windows Recovery Environment.

10. Passkey improvements

Windows 11 now has a redesigned Windows Hello and the ability to integrate third-party password managers into it. The operating system can tap into things like 1Password to use your stored passkeys to sign you into apps, websites, etc. It can also save a new passkey there, for a more seamless password-free experience. As a long-time 1Password user who tries hard to adopt passkeys, it is a very nice and welcome improvement.

These are only 10 features and improvements that stood out among the rest for me. What are your favorite updates to Windows 11 in 2025? Share your thoughts in the comments.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 1 January 2026 at 3:31 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

As a reminder, feature deprecation means Microsoft is no longer actively developing something. Deprecated features may continue working for some time until Microsoft removes them for good.

Deprecated and removed Windows features in 2025

Location History | Deprecated in February 2025

The Location History feature was a Cortana-related API. It was responsible for storing 24 hours of location data, allowing the assistant to tap into it to provide relevant and helpful information. Note that this change does not affect location services, and the operating system can still detect your location for position-based details, such as weather, etc.

Line Printer Daemon | Deprecated with Windows Server 2012, Reminder in February 2025

This printing protocol was deprecated over a decade ago, but was still available. Microsoft reminded users about the upcoming removal so that they can prepare their environments, as it affects the ability to print to a server on UNIX systems. Microsoft recommends switching from LPR/LDP to IPP.

Windows UWP Map Control and Windows Maps Platform API | Deprecated in April 2025

These things allowed developers to implement map controls in their UWP apps. While they continue functioning in existing apps, Microsoft no longer provides updates to the control and API. The company recommends that developers switch to Azure Maps in their applications.

VBS Enclaves on older Windows 11 versions | Deprecated in April 2025

This security feature allows applications to create isolated areas in memory for trusted execution environments (like extra-secure rooms for important tasks). In April, Microsoft deprecated VBS enclaves in Windows 11 version 23H2 and older. In May, it clarified that existing enclaves signed with the EKU 1.3.6.1.4.1.311.76.57.1.15 would continue to be supported for all Windows 11, version 23H2 and earlier versions of Windows, so long as no changes are made to the enclave that require a re-sign of it.

Windows Maps | Deprecated in April 2025

Windows 11 no longer has a stock Maps app. Microsoft deprecated it and removed it from the Microsoft Store in July 2025. Existing installations also no longer work, and Microsoft suggests using web-based Bing Maps.

Device Metadata | Deprecated in May 2025

Device metadata was introduced with the release of Windows 7. It allowed manufacturers to bundle their devices with a bunch of extra details, icons, and logos. Device metadata in XML is no longer supported, and OEMs cannot submit it to the Partner Center. However, Microsoft still allows including additional device information using INF.

Dev Home | Killed in May 2025

Microsoft introduced the Dev Home app in 2023, giving devs a special app for managing certain settings and features on Windows 11 systems. In January 2025, users discovered that Microsoft decided to kill the app in May 2025. The most important parts of the app migrated to the Settings app, so nothing of big value was lost.

Windows PowerShell 2.0 | Deprecated in 2017, Reminder in June 2025

PowerShell 2.0 is quite an old version of the app, and it was deprecated eight years ago. It is no longer bundled with Windows 11 (starting with version 24H2). Scripts that rely explicitly on PowerShell 2.0 will fail to launch, so Microsoft told developers to update them to avoid disruptions.

WMIC | Deprecated in 2016, Reminder in September 2025

WMIC was a command-line interface for Windows Management Instrumentation, a method for IT Admins to manage their systems and retrieve information about hardware, software, services, and more. WMIC was turned off by default in early 2024, and the latest Windows 11 releases no longer have it as an optional component. Microsoft replaced WMIC with Windows PowerShell for WMI.

Legacy Web Components | Deprecated in September 2025

Legacy Web View, Windows 8 UWP/HTML/JavaScript apps (hosted Web Apps), Legacy Progressive Web Apps (PWA), and Legacy Microsoft Edge DevTools (EdgeHTML) are now deprecated. These legacy web components from the EdgeHTML (the original Microsoft Edge) era are still available, but Microsoft warned that they would eventually stop receiving security updates, and they would be removed from Windows. Microsoft strongly recommends switching to WebView2.

Do you think there is something else that Microsoft should remove from Windows 11? Share your thoughts in the comments.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 1 January 2026 at 3:29 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

What is indexing?

In simple terms, indexing is the process of using metadata to structure items in a format that is easily searchable and does not require you to look through each item of your dataset every time. Think of an index similar to the index page/table of contents of a book, which lists each chapter and its corresponding page number. Through this, you can directly jump to the page number of the chapter that interests you rather than scouring through each page manually. Indexing in Windows works the same way on a base level, where it categorizes your on-device items so that they can be searched in a faster manner.

How do I enable indexing in Windows 11?

Windows 11 has search indexing enabled by default. You can navigate to it through Windows Settings > Privacy & security > Search. By default, it is set to Classic mode. In this configuration, Windows only indexes the Documents, Pictures, Desktop, and Music folders. That means that if you store your content in any of these locations, it will appear in search results much faster than other items.

Within Classic mode, you can press the Customize search locations button, which opens the legacy Indexing Options UX from the Control Panel. Here you can manually add other folders that you would like indexed too. This is particularly useful when you store files in directories other than those included by default, and would like those locations to be included in faster search results, too.

However, if you would prefer to have all of your PC's content indexed, you can simply switch to the Enhanced setting. It performs the initial indexing only when your PC is plugged in, and you do have the option to manually exclude some folders.

Great! But why doesn't Microsoft enable enhanced indexing by default?

While Enhanced indexing surfaces search results in Windows faster, there is a reason that it's not the default configuration. This is because indexes take up space. Microsoft indicates that a general rule of thumb is that an index will take up roughly 10% of the size of the overall content being indexed. This means that if your PC has 1TB of content stored on it, an index for it may take up 100GB. That's a lot on a device with limited resources, and customers likely wouldn't be happy if Windows was forcibly reducing their available storage space, which is why it's not enabled by default. The Redmond tech giant even says that sometimes an index takes up more than 10% of available space, so it's clearly not something that should be enforced upon users.

Another thing to keep in mind is that updating an index can take up time and resources on your PC. The initial indexing process itself usually takes a few hours, but even after that, if you delete or move a file, its corresponding index will be updated, so this process is constantly running in the background to monitor any changes. Let's go back to the book analogy, where if a book publisher decides to start their book from page number 2 rather than page number 1, they would also need to increment the page number of each chapter by 1 in their index/table of contents. This continuous potential strain on a PC is also why Enhanced indexing is not the default option, as Classic strikes a good balance between performance and resource utilization.

Oh, anything else that I need to know about indexing?

Indexing in Windows is leveraged by numerous applications, including Windows Search and File Explorer. The index is typically built using all properties of a file, such as file path and file name, but you also have the option to index their internal content, so it's easier to search within them too. All data related to the index is stored locally, and Microsoft promises not to exfiltrate that data into its cloud or to train its AI models. That said, certain third-party apps can use the index too, so make sure that you only install trusted applications.

It is possible to index encrypted files as well, and to toggle between the index of file content versus file properties only. You can configure this through the Advanced indexing options UX on the same Windows Settings page referenced previously.

Copilot+ PCs leverage semantic indexing in addition to traditional indexing. This means that items closely related to each other in meaning may also be surfaced in search results, rather than those that match identically in terms of file properties or lexical similarity. This is done using the on-device model and is enabled by default.

Overall, it's good that Microsoft indexes some directories by default and gives users the option to index more locations or even their entire PC. The latter option does come with some trade-offs, though, so do keep them in mind before you enable Enhanced indexing.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 1 January 2026 at 3:27 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

Starting with enhancements to chat and collaboration, many Teams users will be pleased to know that they can now pop out certain core components like chats, calls, calendar, and more in separate windows by right-clicking on their icons. Additionally, Microsoft is giving team owners more control over who can join a private team using a code by requiring owner approval regardless of the presence of a code. This improves the overall security posture of this particular feature.

Microsoft has also made several improvements to the Interpreter and multilingual meeting experience, where the spoken language in meetings is now automatically detected and propagated across live captions and transcription services. As the Interpreter initializes, it will also show a "preparing" status as a visual indicator before people start speaking.

In addition, the following devices are now Teams-certified:

• EPOS IMPACT 500 MS UC ANC WL USB-C+A Headset
• EPOS IMPACT 400 ANC Headset
• Lenovo Webcams: FHD, QHD and 4K Pro
• Lenovo Dual-mode Wireless ANC Headset 6550 (USB-A and USB-C)
• The Yealink MVC S90
• The Yealink S50
• The Yealink MVC S40

There were some new features introduced for Teams Phone, frontline worker solutions, and Teams fundamentals and security, too. These include the general availability of Microsoft 365 Copilot chat in the Calls app post-calling experience, Compliance Recording ISV Support for Call Queues in Microsoft Teams Phone, Own Domain Impersonation Detection in messages, and the availability of the Frontline Hub in the Teams Admin Center. You can read more about them here.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 1 January 2026 at 3:26 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

About Light Switch

Windows 11 has support for both light and dark modes, and while it continues to gain enhancements from time to time, one configuration sorely lacking in its implementation is the ability to automatically switch themes based on different criteria. Various third-party tools have closed this gap over time, but the recent addition of Light Switch to PowerToys was certainly a very welcome move.

This simple utility says exactly what it says on the tin: it allows you to switch between light and dark themes, depending upon certain conditions. There are many benefits of enabling dark mode during your Windows 11 usage, including less strain on your eyes and better battery life on OLED devices.

Light Switch Functionalities

Light Switch in PowerToys is disabled by default, but once you enable it, you can manually toggle between themes using the Win + Ctrl + Shift + D shortcut, which is configurable, of course. If you manually toggle between modes, it usually takes a couple of seconds before the theme is switched completely.

Furthermore, you can set three modes:

• Off: Automatic switching is disabled
• Fixed hours: Manually adjust the time of the day when you want to enable and disable dark mode
• Sunset to sunrise: You will first need to synchronize your device's location either through location services or manually entering your latitudinal and longitudinal coordinates, and then configure if you want to offset the sunrise or sunset time in case you want to trigger dark mode earlier or later

Additionally, you can choose if you want to apply dark mode only to system elements (Taskbar, Start menu, etc.), supported applications, or both.

Conclusion: Light Switch is an awesome tool, but I wish it was present in Windows 11 by default

Light Switch is such a simple, but useful tool that does exactly what it says on the tin. That said, I do want to emphasize that it should be a default feature of Windows 11 rather than a utility inside PowerToys. I don't think it's intended for power users, as it benefits everyone, so why restrict it to PowerToys? Maybe it will eventually become a part of the OS, just like Snap Layouts graduating from PowerToys to a Windows 11 staple, and I do wish we see it happen with Light Switch too.

This Closer Look piece is based on version 0.96.1 of Microsoft PowerToys, which is the latest version at the time of writing. Functionalities may change in newer versions of PowerToys when they become available.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 1 January 2026 at 4:35 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

About Color Picker

Color Picker is one of the smaller utilities in PowerToys, just like Awake, and while I am not particularly a fan of the latter, as discussed here, I do think Color Picker is one of the better toys available in this package. Color Picker basically allows users to copy the exact color value from their screen and then copy it to the clipboard in a configurable format.

For example, if you activate it using the default Win + Shift + C shortcut, your cursor will get a small tooltip next to it, which will show you the hex value of the color of the element that you are currently hovering on. When you click on an element, you'll be greeted with a separate window that shows you the hex, RGB, and HSL (Hue, Saturation, Light) values of the selected color, which you can copy or configure further using the dropper options at the top or the color bar on the left. This makes it an excellent tool for designers and developers who are building their own products and want to reference a color scheme without using other third-party tools.

Color Picker Functionalities

As with most PowerToys utilities, you can change the activation shortcut for Color Picker as well as the trigger behavior, where you can choose between directly opening the editor or picking a color first. I personally prefer the latter option. You can also choose what happens if you right-click, left-click, or middle-click on your mouse.

In addition, you can define which format gets copied to the clipboard by default; I usually keep it as hex, but you can set it to other formats like CMYK, HSL, RGB, HSV, etc. You can also add these formats directly to the editor and arrange them by order of preference. In addition, you can define a custom color format as well, but this option is more geared towards advanced users.

Meanwhile, the editor UX in Color Picker may appear slightly complicated at first, but it is quite powerful too. It stores a history of the most recent 20 colors, shows you the value of the currently selected color, enables you to fine-tune the selected color, and highlights similar color options. You can manually copy values of different color formats, export them, and remove colors from the history, too.

Since the color value depends highly upon the cursor position, animated elements that show different colors based on hover state can be difficult to capture. Microsoft recommends following the process below to capture the non-hover state of such elements:

• Move the mouse pointer close, but not over the element.
• Zoom in by scrolling the mouse wheel up. The image will be frozen.
• In the enlarged area, you can pick the color of the element.

Conclusion: Color Picker is a winner for designers and developers

Although Color Picker doesn't appeal to the mass market, it does offer a handy utility for the right type of user, namely, designers and developers. It is very easy to use and understand, and offers a great deal of configurability in UX and color formats. It's also nice to see that it bundles relatively niche color formats for those who need them too.

That said, Color Picker does have a few noteworthy limitations. It cannot display on top of the Start Menu or the Action Center, and does not support Wide Color Gamut (WCG) and High Dynamic Range (HDR) formats. Additionally, if you want to see the color values of an application running in administrator mode, you'll need to first run PowerToys with administrator privileges too. However, I don't think these are major limitations, and Color Picker should meet the needs of most users anyway.

This Closer Look piece is based on version 0.96.1 of Microsoft PowerToys, which is the latest version at the time of writing. Functionalities may change in newer versions of PowerToys when they become available.

You can check out our other Closer Look pieces for PowerToys utilities here:

• Closer Look: Advanced Paste in Windows PowerToys
• Closer Look: Awake is not my favorite tool in PowerToys on Windows
• Closer Look: Command Palette in PowerToys puts Windows Search to shame

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 1 January 2026 at 4:33 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025: 5,700+

RIP Matrix

2025 was a rough year for Microsoft, at least when it comes to public perception. The tech giant was accused of pushing the "enshittification of Xbox, Surface, and even Windows itself" on purpose. Microsoft also drove away its defenders.

Windows 11, Microsoft's flagship OS, had a disastrous year full of bugs, half-baked AI features, and other issues.

While there's plenty of justification for doom and gloom, there are some good things on the horizon. An improved dark mode and some Microsoft Teams features won't be enough to fix what's wrong in Redmond, Washington, but there are some genuinely exciting changes on the way.

Love: Qualcomm Snapdragon X2 support

The last 12 months have been an incredibly frustrating time for Windows fans. For the first time in a long while, it feels like Windows is suffering from a lack of focus from the people at the top.

Support for Windows 10 ended in October, and this year was the perfect time to strengthen Windows 11 as a viable replacement for millions of users. Instead, Microsoft spent most of it shoving the OS full of half-baked AI features, all while letting the quality bar slip and shipping new bugs and issues on an almost monthly cadence.

Everything Microsoft has done when it comes to Windows this year has eroded the platform's reputation in ways that I haven't seen since Windows 8. Today, it feels like people hate Windows 11 with a passion, much more so than they did when 2025 first started.

There are so many problems with Windows as a platform right now that it's hard to know where to begin.

AI obsession is hurting Windows

If you’ve been following our coverage for the last few years, you’ll already know that 2025 is the year that Windows 10 died. Technically.

“Died,” because Microsoft’s formal end-of-support date came and went on October 14, as the company had been saying for years. “Technically,” because it’s trivial for home users to get another free year of security updates with a few minutes of effort, and schools and businesses can get an additional two years of updates on top of that, and because load-bearing system apps like Edge and Windows Defender will keep getting updates through at least 2028 regardless.

But 2025 was undoubtedly a tipping point for the so-called “last version of Windows.” StatCounter data says Windows 11 has overtaken Windows 10 as the most-used version of Windows both in the US (February 2025) and worldwide (July 2025). Its market share slid from just over 44 percent to just under 31 percent in the Steam Hardware Survey. And now that Microsoft’s support for the OS has formally ended, games, apps, and drivers are already beginning the gradual process of ending or scaling back official Windows 10 support.

Windows 10 is generally thought of as one of the “good” versions of Windows, and it was extremely popular in its heyday: the most widely used version of Windows since XP. That’s true even though many of the annoying things that people complain about in Windows 11 started during the Windows 10 era. Now that it’s time to write Windows 10’s epitaph, it’s worth examining what Microsoft got right with Windows 10, how it laid the groundwork for many of the things people dislike about Windows 11, and how Microsoft has made all of those problems worse in the years since Windows 11 first launched.

Windows 10 did a lot of things right

Like Windows 7, Windows 10’s primary job was to not be its predecessor. Windows 8 brought plenty of solid under-the-hood improvements over Windows 7, but it came with a polarizing full-screen Start menu and a touchscreen-centric user interface that was an awkward fit for traditional desktops and laptops.

And the biggest thing it did to differentiate itself from Windows 8 was restore a version of the traditional Start menu, altered from its Windows XP or Windows 7-era iterations but familiar enough not to put people off.

Windows 10 also adopted a bunch of other things that people seemed to like about their smartphones—it initially rolled out as a free upgrade to anyone already running Windows 7 or Windows 8, and it ran on virtually all the same hardware as those older versions. It was updated on a continuous, predictable cadence that allowed Microsoft to add features more quickly. Microsoft even expanded its public beta program, giving enthusiasts and developers an opportunity to see what was coming and provide feedback before new features were rolled out to everybody.

Windows 10 also hit during a time of change at Microsoft. Current CEO Satya Nadella was just taking over from Steve Ballmer, and as part of that pivot, the company was also doing things like making its Office apps work on iOS and Android and abandoning its struggling, proprietary browser engine for Edge. Nadella’s Microsoft wanted you to be using Microsoft products (and ideally paying for a subscription to do so), but it seemed more willing to meet people where they were rather than forcing them to change their behavior.

That shift continued to benefit users throughout the first few years of Windows 10’s life. Developers benefited from the introduction and continuous improvement of the Windows Subsystem for Linux, a way to run Linux and many of its apps and tools directly on top of Windows. Microsoft eventually threw out its struggling in-house browser engine for a new version of the Edge browser built on Chromium—we can debate whether Chromium’s supremacy is a good thing for an open, standard-compliant Internet, but switching to a more compatible rendering engine and an established extension ecosystem was absolutely the more user-friendly choice. Both projects also signaled Microsoft’s growing engagement with and contributions to open-source projects, something that would have been hard to imagine during the company’s closed-off ’90s and ’00s.

Windows 10 wasn’t perfect; these examples of what it did right are cherry-picked. But part of the operating system’s reputation comes from the fact that it was originally developed as a response to real complaints and rolled out in a way that tried to make its changes and improvements as widely accessible as possible.

But Windows 10 laid the groundwork for Windows 11’s problems

As many things as Windows 10 did relatively well, most of the things people claim to find objectionable about Windows 11 actually started happening during the Windows 10 era.

Right out of the gate, for example, Windows 10 wanted to collect more information about how people were using the operating system—ostensibly in the name of either helping Microsoft improve the OS or helping “personalize” its ads and recommendations. And the transition to the “software-as-a-service” approach helped Windows move faster but also broke things, over and over again—these kinds of bugs have persisted on and off into the Windows 11 era despite Microsoft’s public beta programs.

Windows 10 could also get pushy about other Microsoft products. Multiple technologies, like the original Edge and Cortana, were introduced, pushed on users, and failed. The annoying news and weather widget on the taskbar was a late addition to Windows 10; advertisements and news articles could clutter up its lock screen. Icons for third-party apps from the Microsoft Store, many of them low-rent, ad-supported time-waster games, were added to the Start menu without user consent. Some users of older Windows versions even objected to the way that the free Windows 10 upgrade was offered—the install files would download themselves automatically, and it could be difficult to make the notifications go away.

Even the mandatory Microsoft Account sign-in, one of the most frequently complained-about aspects of Windows 11, was a Windows 10 innovation—it was easier to circumvent than it is now, and it was just for the Home edition of the software, but in retrospect, it was clearly a step down the road that Windows 11 is currently traveling.

Windows 11 did make things worse, though

A desktop with an extra informational wallpaper icon, a Start menu full of apps you may or may not use, and a

busy taskbar will greet you the first time you log in to Windows 11.

Andrew Cunningham

 

A cleaner desktop, taskbar, and Start menu, with many extraneous elements removed. With the exception of Edge,

most Windows apps can be retained or removed as you prefer.

Andrew Cunningham

But many of Windows 11’s annoyances are new ones. And the big problem is that these annoyances have been stacked on top of the annoying things that Windows 10 was already doing, gradually accumulating to make the new PC setup process go from “lightly” to “supremely” irritating.

The Microsoft Account sign-in requirement is ground zero for a lot of this since signing in with an account unlocks a litany of extra ads for Microsoft 365, Game Pass, and other services you may or may not need or want. Connecting to the Internet and signing in became a requirement for new installations of both the Home and the Pro versions of Windows 11 starting with version 22H2, and while workarounds existed then and continue to exist now, you have to know about them beforehand or look them up yourself—the OS doesn’t offer you an option to skip. Microsoft will also apparently be closing some of these loopholes in future updates, making circumvention even more difficult.

And if getting through those screens when setting up a new PC wasn’t annoying enough, Windows 11 will regularly remind you about other Microsoft services again through its Second Chance Out-Of-Box Experience screen, or SCOOBE. This on-by-default “feature” has offered to help me “finish setting up” Windows 11 installations that are years old and quite thoroughly set up. It can be turned off via a buried checkbox in the Notifications settings, but removing it or making it simpler to permanently dismiss from the SCOOBE screen itself would be the more user-friendly change, especially since Microsoft already bombards users with “helpful reminders” about many of these same services via system notifications.

Microsoft’s all-consuming pivot to generative AI also deserves blame. Microsoft’s Copilot push hasn’t stopped with the built-in app that gets a position of honor on the default taskbar—an app whose appearance and functionality have completely changed multiple times in the last couple of years as Microsoft has updated it. Microsoft changed the default Windows PC keyboard layout for the first time in 30 years to accommodate Copilot, and Copilot-branded features have landed in every Windows app from Word to Paint to Edge to Notepad. Sometimes these features can be uninstalled or turned off; sometimes they can’t.

It’s not just that Microsoft is squeezing generative AI into every possible nook and cranny in Windows; it’s that there seems to be no feature too intrusive or risky to make the cutoff. Microsoft nearly rolled out a catastrophically insecure version of Recall, a feature for some newer PCs that takes screenshots of your activity and records it for later reference; Microsoft gave its security an overhaul after a massive outcry from users, media, and security researchers, but Recall still rolled out.

The so-called “agentic” AI features that Microsoft is currently testing in Windows come with their own documented security and privacy risks, but their inclusion in Windows is essentially a foregone conclusion because Microsoft executives are constantly talking about the need to develop an “agentic OS.” There’s a fine line between introducing new software features and forcing people to use them, and I find that Microsoft’s pushiness around Windows 11’s AI additions falls on the wrong side of that line for me pretty much every single time.

Finally, while Windows 10 ran on anything that could run Windows 7 or 8, Windows 11 came with new system requirements that excluded many existing, functional PCs. The operating system can be installed unofficially on PCs that are several years older than the official cutoff, but only if you’re comfortable with the risks and you know how to get around the system requirements check.

I find the heightened requirements—implemented to improve security, according to Microsoft—to be more or less defensible. TPM modules enable seamless disk encryption, Secure Boot protects from threats that are otherwise invisible and hard to detect, and CPU makers like Intel and AMD only commit to supporting older processors with firmware-level security patches for so long, which is important in the era of hardware-level security exploits.

But the requirements don’t feel like something Microsoft has imposed to protect users from threats; they feel like something Microsoft is doing in order to upsell you to a new PC. Microsoft creates that impression when it shows Windows 10 users full-screen ads for new Copilot+ PCs, even when their systems are capable of upgrading to the new operating system directly. People are already primed to believe in “planned obsolescence,” the idea that the things they buy are designed to slow down or fail just in time to force them to buy new things; pushing people to throw out functioning PCs with full-screen ads does nothing to dispel this notion.

Windows 11 could still be great

I still believe that Windows 11 has good bones. Install the Enterprise version of the operating system and you’ll get a version with much less extra cruft on top of it, a version made to avoid alienating the businesses that pay good money to install Windows across large fleets of PCs. Microsoft has made huge strides in getting its operating system to run on Arm-based PCs. The Windows Subsystem for Linux is better than it’s ever been. I’m intrigued by the company’s efforts to make Windows a better operating system for gaming handhelds, Microsoft’s belated answer to Valve’s Steam Deck and SteamOS.

But as someone with firsthand experience of every era of Windows from 3.1 onward, I can say I’ve never felt as frustrated with the operating system as I have during Windows 11’s Copilot era. The operating system can be tamed with effort. But the taming has become an integral part of the new PC setup process for me, just as essential as creating the USB installer and downloading drivers and third-party apps. It’s something my PC needs to have done to it before it feels ready to use.

Windows 10 was far from perfect. But as we mark the first stage of its multi-year passing, it’s worth remembering what it did well and why people were willing to install it in droves. I’d like to see Microsoft recommit to a quieter, cleaner version of Windows that is more willing to get out of the way and just let people use their computers the way they want, the same way the company has tried to recommit to security following a string of embarrassing breaches. I don’t have much hope that this will happen, but some genuine effort could go a long way toward convincing Windows 10-using holdouts that the new OS actually isn’t all that bad.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 30 December 2025 at 4:34 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

With Files 4.0.24, you can now preview files by simply pressing Spacebar. This convenient feature has been a staple feature of Mac's Finder, and this year, Microsoft finally brought it to Windows 10 and 11 with a PowerToys update. Just keep in mind that you need to have PowerToys installed alongside Files to make use of that upgrade.

In addition to integrating PowerToys Peek, the latest Files update has a few smaller quality-of-life improvements that make the app a little more convenient and useful. For example, right-clicking items in the sidebar now has the "Copy" action, and creating a new document automatically pre-fills its name, such as "New Text Document" or similar. The latter lets you quickly create documents and files without wasting time on renaming them. Omnibar now casts shadows, the Status Center now properly wraps longer text in folder names, and tag search received support for AND/OR operators.

Finally, here are some of the fixed bugs:

• Fixed a crash that would occur when resizing the window in Dual Pane mode.
• Fixed a crash that would occur when switching git branches during a merge conflict.
• Fixed an issue where PNG-based .ico files with black backgrounds displayed incorrectly in the sidebar.

You can download Files for Windows 10 and 11 from the official website, its GitHub repository, or the Microsoft Store. The latter option is paid—developers use it as a convenient method for users to support them. Otherwise, the app is completely free.

Full changelog for Files 4.0.24 is available here.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 30 December 2025 at 4:31 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

The majority of the changes in this release are concentrated in the drivers subsystem. There are significant documentation fixes in kernel-doc for the Snapdragon driver and various reverts in the Display Processing Unit Code. The update also contains solutions for numerous quirks on specific laptops from Asus, HP, Honor, and Medion, as well as resource leak fixes in probe paths. There are also updates for Intel and NXP/Freescale audio components.

The third release candidate also brings several reverts in the Type-C UCSI driver, fixes for resource leaks in various controller drivers including dwc3, renesas, and lpc32xx, and storage/phy adjustments. The update also brings fixes for device initialization, sysfs information restoration, and architecture bug fixes for Longson, Xilinx, and Renesas.

Architecture specific updates include the enabling of CPU idle state detection for POWER11, updated PMU tests, and fixed interrupt/syscall exit behaviours on 32-bit systems. For RISC-V systems, ISA extension descriptions for Zilsd and Zclsd have been added, there is improved signal handling, and atomic operations for non-SMP configurations have been optimized.

We are now around early to halfway through the 6.19 kernel cycle, which should last seven or eight weeks. Once it has been released, Linux distributions will make it available after a period of testing. Some distributions won’t release it if they stick to LTS kernels, but if you’re running something Arch-based or Fedora, you’ll likely get it soon after released.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 30 December 2025 at 4:31 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

 ZDNET's key takeaways

• The Linux desktop will continue to grow.
• Linux and open-source security will both improve.
• Firefox is in deep, deep trouble.

Clearly, AI will play a larger role in Linux and open source next year, but that's true of pretty much all technology. However, while AI will be used to help develop the Linux kernel, no one is predicting, a la Windows, that AI will be used to rewrite the entire codebase by 2030. That said, open source will remain at the heart of AI.

The continued rise of the Linux desktop

For Linux, I see more desktop distributions aimed at ex-Windows users. As a result, the Linux desktop, which has already been growing faster than ever, will continue to grow even faster. Microsoft is helping by continuing to push AI down Windows users' throats. Attention Satya Nadella, Microsoft's CEO: Windows fans don't want AI. Sure, some do. Most, however, don't.

In addition, as my colleague on the Windows side of life, Ed Bott, points out, Microsoft's current course toward more restrictions on which applications Windows can run, and a monthly subscription model for Windows, is almost certain to disillusion even hardcore Windows users.

All this means Linux's opportunity for growth will only continue to grow. Now, we'll see if Linux can take advantage of its new chances. You see, as Linus Torvalds himself pointed out long ago, we have too many Linux desktops.

Imagine yourself at a grocery store in another country, and you want a new breakfast cereal. You're presented with over a hundred different boxes, and you don't know a thing about any of them. How will you decide? Or will you just throw up your hands and order Windows 12 via Amazon from back home because you can't make heads or tails of all the local cereals?

It's the same with Linux. If you go to DistroWatch, you'll find over a hundred desktops. I can't tell them all apart, and I cover this stuff for a living. We need a distro to step forward to become the top choice.

That has never happened because every Linux distribution maker thinks their way is the best way. The closer you look, the more differences you'll find between the distros. For example, there are more than half a dozen viable Linux desktop interfaces such as KDE Plasma, GNOME, Cinnamon, etc.

I have some ideas on how to address this issue, but that's a story for another day. For now, even though there are still too many Linux desktops, the Linux desktop will continue to grow.

Rust becomes 'normal' in kernel and core tools

Earlier this month, Linux kernel developers formally ended the "Rust experiment," declaring Rust a permanent core language for Linux. Indeed, the Direct Rendering Manager (DRM) graphics maintainers are already talking about requiring Rust for new drivers within about a year.

Even before, the Debian Linux developers had decided that, by May 2026, all further development of its core APT package manager would be done in Rust. Why? Because Rust is memory safe. This stops in their tracks many kinds of security problems that C, Linux's main language, has proven all too prone to over the decades.

Don't think, however, that it's all peaches and cream now for Rust on Linux. 

As Miguel Ojeda, leader of the Rust for Linux project, has said, "Rust is here to stay, ... this does not mean that everything works for every kernel configuration, architecture, toolchain, etc." 

There's still a lot of work to be done. That said, Ojeda also noted that Android 16 devices based on the 6.12 Linux kernel include the new Rust-based anonymous shared memory allocator (ashmem). That means millions of devices are already using Rust for Linux in production. Rust is here. Rust is real. And Rust is already at work.

However, unlike Microsoft, which appears to plan to move all of Windows to Rust by 2030, there are no such plans for Linux. Someday, all of Linux may be written in Rust, but I can't see it happening until the 2050s, and maybe not even then. After all, for pure performance, short of writing code in assembler, nothing's faster than C.

Oh, and for Windows' Rust plans? Good luck with converting that mess of a codebase into Rust in four years. Fourteen? Maybe.
Immutable Linux goes mainstream

Immutable Linux distributions are gaining traction because read-only system images, atomic updates, and transactional package layers significantly simplify rollback and reduce "dependency hell." Analysts are explicitly framing immutability as a "new era of security and stability."

Now, immutable Linux distros, such as Fedora Silverblue, OpenSUSE MicroOS, and Ubuntu Core, have been around for a while. Previously, though, only Linux enthusiasts used them in containers. Enterprise Linux is now switching to them.

The leading example of an immutable business is the new Red Hat Enterprise Linux (RHEL) 10. While you can still use the old tools to maintain RHEL, you now have the choice of going the immutable route. As time goes on, I expect more enterprise distributions to follow this trend. It's much easier to manage.

Security and supply-chain hardening across the stack

Some inherent changes, such as Rust's integration into Linux, are guaranteed to make Linux more secure. 

Linux and open source are also on track to become more secure in 2026, as the ecosystem simultaneously hardens the kernel, professionalizes supply-chain defenses, and scales coordinated security programs like the Open Source Security Foundation (OpenSSF) across vendors and governments.

Starting with Linux, the Kernel Self-Protection Project and similar efforts are pushing more exploit-mitigation features upstream. Simultaneously, risk-based patching and AI-assisted triage are making it easier to prioritize and deploy fixes at scale.

At the same time, open-source supply-chain attacks have driven widespread adoption of Software Bill of Materials (SBOMs), Supply Chain Levels for Software Artifacts (SLSA), and signed provenance. 

Sigstore is now integrated into major platforms like GitHub and GitLab, which will make verifiable signing routine by 2026. That's important because, more than ever, developers need to know what's actually in any given program and who really wrote it.

It's not just, by the way, that this just makes good security sense. In Europe, with the EU Cyber Resilience Act (CRA) now in force, it's the law. The CRA requires anyone, by the end of next year, who sells any product containing software, including open-source programs, to have an SBOM. Period. End of statement.

Open source becomes essential for AI agents

Last year, over in AI land, everyone was talking about how 2025 would be the year of Agentic AI. Leading the hype pack was OpenAI founder Sam Altman, who said, "In 2025, we may see the first AI agents join the workforce and materially change the output of companies." He was wrong. Agents aren't ready yet.

If AI agents are really going to be all that -- and I'm not convinced they will be -- it will be because of the open-standard Agentic AI Foundation (AAIF). This new, open standards industry-backed consortium gets pretty much everyone who's anyone in AI agents on the same page for sharing data and making interactions easy for agents. Without interoperability, agents will go nowhere.

Firefox falls

I started this story by mentioning how people are annoyed with Microsoft for forcing AI on its users. Their bad temper is nothing -- nothing -- compared to how Firefox users are reacting to Mozilla pushing AI into their favorite browser.

When Mozilla declared that AI would be Firefox's future, its users quickly said, and I quote from the Mozilla Connect list, "Once again, Mozilla is SPRINTING to chase after the stupidest tech-brained trends and not actually focused on improving the product at all." 

This was followed up by comments such as "Stop chasing slop and actually work on making sure your browser competes with Chromium in REAL areas like speed and functionality, not in digital hallucinations." And one I especially liked: "I would LOVE a function that helps me remove AI slop from my everyday browser. I would really encourage Mozilla to develop this."

Mozilla backed off quickly. 

On Mastodon, Mozilla retreated to "Firefox will have an option to completely disable all AI features. We've been calling it the AI kill switch internally. I'm sure it'll ship with a less murderous name, but that's how seriously and absolutely we're taking this."

I think the damage has already been done. As I've been saying for years, Firefox has been declining. 

The people in charge of the project keep chasing one new tech craze after another, such as briefly accepting cryptocurrency donations and Firefox Hello, a built-in WebRTC video and voice calling feature, while abandoning popular features such as built-in RSS support (Live Bookmarks) and legacy add-ons/XUL extensions. 

Mozilla has also pursued such massive projects as its own device operating system, Firefox OS; a Virtual Private Network (VPN); and Mozilla Monitor Plus, a paid data-broker scan and removal service. What Mozilla hasn't been doing, many Firefox users have complained, is improving Firefox's performance.

The result has been that Firefox's popularity has been sinking into irrelevance for years. Over the last 90 days, Firefox's market share has sunk in the US to a mere 1.7%. At one time, Firefox had a 34.1% market share. It will never see those numbers again. I fear, by this time next year, thanks to how annoyed its most loyal users are, it will drop below 1%.

Firefox was once one of open source's greatest success stories. Now, its day is almost done.

Well, it couldn't all be great days ahead, could it? Still, taken all in all-in-all, I expect 2026 to be a banner year for both Linux and open-source software. 

Source

Although most people download the free Community edition (as they should), the paid versions of the IDE do offer some capabilities not present in the free variant, and it's worth knowing about them, especially in professional scenarios.

A summary of the comparison can be seen below, with more details that follow:

• Target market: Visual Studio 2026 Community is intended for individuals and non-enterprise customers only. Professional and Enterprise are meant for organizational use. Microsoft defines an enterprise organization as an entity having over 250 users and at least $1 million in annual revenue.
• Development platforms: All three flavors are identical, but Embedded Assemblies are supported only in Visual Studio Enterprise.
• Features: CodeLens, Peek Definition, Refactoring, One-Click Web Deployment, Model Resource Viewer, and Multi-Targeting are available across all variants. However, Live Dependency Validation, Architectural Layer Diagrams, Architectural Validation, and Code Clone are only available in Visual Studio Enterprise. Additionally, this is the only flavor that has full support for Dependency Graphs and Code Maps, whereas Community and Professional can open diagrams made in older Visual Studio editions in read mode only.
• Debugging, diagnostics, and testing: All three editions support Code Metrics, Graph Debugging, Static Code Analysis, Performance and Diagnostics Hub, Code Coverage, and Unit Testing. However, Tier Interaction Profiling, IntelliTrace, Code Map Debugger Interaction, .NET Memory Dump Analysis, Live Unit Testing, and IntelliTest are exclusive to Visual Studio Enterprise.
• Collaborative tools: All flavors are identical.
• Pricing: As mentioned previously, Visual Studio Community can be downloaded free of charge, Professional Monthly costs $45/user/month, and Enterprise Monthly costs $250/user/month. That said, Microsoft offers bundles called Professional Standard ($99.99/user/month) and Enterprise Standard ($499.92/user/month), which offer some more advanced tooling associated with development. You can check out additional details here.

All in all, Community and Professional are pretty much the same, but the latter is meant for organizational use as it also offers the Basic Plan for Azure DevOps. Meanwhile, Enterprise is the beefier edition that contains pretty much every feature that Microsoft has developed so far, including the Basic + Test Plan for Azure DevOps. That said, the Professional Standard and Enterprise Standard are slightly harder sells as they bundle a lot of tools, but at massive costs per user that need to be paid annually.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Sunday 28 December 2025 at 12:12 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

Today, we'll be taking a look at the second toy in the System Tools section in PowerToys, and that is Awake.

About Awake

Windows already has a set of configurations that allows you to set your PC's sleep state through various power plans. While they can fit most use-cases, PowerToys naturally extends this functionality a bit through Awake, which offers primarily time-based controls around your device's sleep state.

That said, Awake does come with some limitations. It can only work if you have signed in to your PC and have enabled the toy, which means that it doesn't work on the lock screen for security purposes. If your device goes to the lock screen at any point, the PC will revert to the default power plan. It's also worth keeping in mind that Awake does not modify your existing power plan; it only overrides it through a set of customized background threads that get terminated once a particular criterion is completed.

PowerToys Awake can be accessed from the taskbar and the PowerToys app, and is enabled by default.

PowerToys Awake Functionalities

Awake basically has four possible configurations. These are described below:

• Disabled (Passive): It keeps utilizing the configured Windows power plan and does not override any settings.
• Keep awake indefinitely: Your PC will always stay awake unless you disable the application or put your PC to sleep intentionally.
• Keep awake for a time interval: Define the time in hours and minutes for how long you want your PC to stay awake. Once this time limit is hit, Awake will revert to its disabled state.
• Keep awake until expiration: Similar to the mode described above, but you can set an exact date and time, rather than configuring a timer.

Additionally, any external displays connected to your PC will still turn off regardless of your PC following one of Awake's modes, unless you toggle on the Keep screen on functionality. Furthermore, if you right-click the Awake icon on the system tray in the Taskbar, you'll be able to identify the currently enabled mode.

PowerToys Awake goes a bit beyond the GUI as well, unlike Advanced Paste. You can configure it as a standalone application that can be triggered through the Command Line Interface (CLI) as described here, or modify the settings.json file as detailed here.

Conclusion: Awake is nice, but I can't really find a use for it

Now, there's no doubt that PowerToys Awake offers a couple of configurations beyond the scope of the power plans built into Windows, but I personally haven't been able to figure out a use case where I would actually use them. In all my personal and organizational workflows, I haven't encountered scenarios where I would have to keep my PC awake for a specific time interval or a set date and time. I configure my power plan such that my PC is always awake, and when I do want to put it to sleep, I do so manually. The available power plans in Windows don't impact my productivity and fit my use cases perfectly. Of course, that doesn't mean that Awake is useless; the reason it's still included in PowerToys is that some people obviously have a use for it. I am just not one of them.

You can check out Awake by downloading PowerToys from the Microsoft Store here.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Sunday 28 December 2025 at 4:00 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

With the software approaching version 1.0 slowly (currently on version 0.96.1), we thought it would be a good time to restart our dormant but popular Closer Look series to walk through PowerToys' collection of 25+ handy utilities. We'll be starting with Advanced Paste today.

About Advanced Paste

Although Windows 11 already offers a very useful Clipboard History feature by default, which is miles ahead of the implementations present in older versions of the operating system, Advanced Paste goes beyond that by adding more formatting options and integration with AI capabilities. It is clubbed under the System Tools category in PowerToys, along with a handful of other utilities. Advanced Paste is a powerful tool in the hands of those who need advanced clipboard management functionalities.

Advanced Paste Functionalities

Advanced Paste is toggled on by default when you install Microsoft PowerToys, but only with offline capabilities enabled - more on AI capabilities towards the end. You can trigger it through the Windows + Shift + V shortcut, which is configurable with some constraints.

In terms of general functionalities, you can toggle various configurations such as clipboard preview, clipboard history, preview of custom format (for AI outputs), and closing of the Advanced Paste window immediately as it loses focus.

Advanced Paste capabilities can further be divided into three distinct categories, discussed below:

• Advanced Text Paste: There are three options available by default for text formatting, namely paste as plain text, markdown, or JSON. These can be very useful in many scenarios, including formatting from XML to JSON and HTML to markdown, which means that the module can be a great utility in the arsenal of programmers. All of these have shortcut keys (Ctrl + 1, 2, 3) in the Advanced Paste window, but you can set shortcut keys to directly trigger these commands without opening Advanced Paste. That's not all, you can paste text as files with auto-generated file names too. Right now, there is support for .txt and .html file formats. You can enable these formats within Advanced Paste's settings and assign them shortcuts, too.
• Advanced Image Paste: As the name suggests, this offers image pasting capabilities, but only a couple right now. You can assign shortcuts to extract text from images and paste it, or even save different file formats (such as .bmp) in PNG format.
• Transcoding to Audio or Video: This is a relatively niche feature set in Advanced Paste, but it has a couple of powerful, offline capabilities in tow. If you have a video file on your clipboard, you can utilize the transcode to .mp3 option to extract the audio from it and save it as an MP3 file, and you can even transcode the whole video to .mp4 (H.264/AAC) format. However, since these are relatively lengthy processes, you can view their progress and choose to cancel them mid-way.

It is worth noting that none of the features mentioned above require AI, and run completely locally, so if you're privacy-conscious, all of these should work out great for you.

AI Integration in Advanced Paste

With the rise of generative AI, it was inevitable that some of that technology would be infused into Advanced Paste. However, the good thing to note here is that it is completely optional, and you do have the ability to choose between online models versus a local model hosted in Foundry Local or Ollama.

If you enable AI features from within Advanced Paste settings, you'll be prompted to add a model provider. Cloud-based options, which require an API key, include OpenAI, Azure OpenAI, Mistral, Google, and Azure AI Inference. You can read more about the integration process here.

Enabling this integration will unlock several new text-pasting options powered by AI. You can input natural language prompts such as "Format this text so each sentence is on a separate line and the writing style should match that from the 1800s, and watch as the magic unfolds. Other possible actions include summarization of text, generating code based on the description of a programming problem in the clipboard history, formatting, and translation, among many other things. If you opt for OpenAI as your model provider, you will also be able to leverage even more advanced Semantic Kernel capabilities, such as chaining of multiple actions, production of new file formats, and more options for image handling.

That said, it is very important to keep in mind that Microsoft does not guarantee the correctness of output in AI scenarios since they are highly dependent on the input format and the model being used. As such, use these only at your own risk and also remember that it can exhaust your API quota depending upon how heavily you lean on any of the aforementioned functionalities.

Conclusion: Advanced Paste is a Beast

As can be summarized from the above, Advanced Paste is a beast that goes beyond a typical clipboard management tool. It can replace many other online and offline tools or workarounds that you utilize for content pasting in different formats, and it is a godsend for many customers, such as programmers. That said, some may also argue that it veers away from the clipboard management territory at times, especially when it comes to AI integration, which expands its feature set by quite a lot. However, it is good that all of that is completely optional, and Advanced Paste works just fine for many use cases within offline boundaries as well.

You can check out Advanced Paste by downloading PowerToys from the Microsoft Store here.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Saturday 27 December 2025 at 4:15 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

We’re ending 2025 with a mixed bag of issues.

To be on the safe side, I’m lowering the MS-DEFCON level just one notch, to 3. Even though the holiday season offers many users the time to apply updates, this year it’s not quite as quiet. Use caution.

For some Windows 11 24H2 and 25H2 users, I’m seeing no side effects from this month’s KB5072033 release. Others had trouble installing the patch on new PCs or on those that had recently been refreshed from scratch. Reader Randy Hollenbeck pointed out a post regarding some Dell PCs:

The problems include failed installations, frequent system freezes, and a non-functional Start Menu.

The culprit in one case was an Intel Arc graphics card where a BIOS setting caused the issue. Disabling the Enable adaptive c-States for discrete graphics option under Advanced Setup | Performance in the BIOS stopped the system crashes.

Another weird side effect is fans that run when the machine is idle and under no load. There are some other odd effects. That’s why I say this is a mixed bag.

If you are in a business setting and need to install this month’s updates, do so — and check carefully for side effects. Consumers — especially anyone with a new PC — may want to hold back on this month’s updates.

I’ve not yet seen Microsoft officially acknowledge these strange issues, but there are enough posts around the Web to convince me that these are not one-offs. One possibility, unconfirmed, is that the underlying issue is a missing driver. Another is a problem with the update installation process itself. Keep in mind that OEMs can create their own components and include them with their build of Windows. That’s why a problem might be seen on one manufacturer’s machines but not on others.

I wish there were some sort of website that I could send you to that would scan your system and  tell you that the side effects I’m describing applied in your case. I test updates on my office and home PCs, and so far I’ve seen no problems. It’s just that there are so many reports of problems that I must urge caution, even though my environment is in the clear. And, of course, there is always the possibility that media coverage is overblown, based on a few isolated reports, and turned into clickbait.

Consumers

The good news for subscribers of the Windows 10 ESU is that there are almost no side effects. I say “almost” because one known issue generated an out-of-band release to address the side effects of the December updates. Out-of-band means it’s an extra important update that is being released outside the normal Patch Tuesday cadence. It used to mean there was a worm or other severe security event that needed to be addressed. Now it’s more likely that it’s fixing some bug introduced by a security patch.

Microsoft Message Queuing (MSMQ) functionality is not often seen in consumer or small businesses but can be seen in Enterprises. Patches for Windows 10 ESU were released on December 18 to fix the issue. I don’t expect consumers to see this, but if you see error messages such as the ones below, you might be affected anyway.

If you are seeing anything like those messages in your event logs, then install KB5074976 from the Microsoft Catalog site as noted in my forum post about MSMQ. You will have to manually click on the download to install it. But again, I’m providing this information in the consumer section only on the off chance that a reader has some unusual application installed on their Windows 10 system that would be using MSMQ. However, this is highly unlikely.

In researching patch issues this week, I did come across an interesting issue with an even more interesting resolution. Windows updates installed on a computer are a combination of two parts: the security updates and the servicing-stack update. Even though servicing-stack updates can’t be uninstalled, they don’t introduce issues and bugs into the system. Security updates are not so clean. It should be possible to uninstall them.

If you’ve been around computers for a while, you are probably used to seeing instructions to use wusa.exe with the uninstall switch. However, this no longer works with Windows 11. Instead, use DISM commands or the GUI to uninstall the update. If either method fails, try this trick that I spotted on a Microsoft support venue saying it had worked for them:

If you have Windows Sandbox enabled, that is the reason. You have to disable it in the “Turn Windows Features on and off” menu.

To do this, press the Windows key and start typing “Turn Windows features … .” Search will quickly find the Windows control panel, probably with just the word “turn.” Click the Control Panel app that shows up, scroll down until you see “Windows Sandbox,” and uncheck it (Figure 1).

Figure 1. Turn off the Windows Sandbox in Settings.

If it was checked, then reboot the PC and try uninstalling the update from the Windows update history window. Ideally, you’ve been keeping backups and can roll your machine back to right before you installed any update.

Businesses

I hope you are not planning to roll out the KB5072033 update at this time. Enough people report having issues with installing this update on brand-new systems. Because I did not have issues on any of my business systems, my thought is that this problem must be in the latest builds, not from corruption occurring on existing systems — even though the error code 0x800f093 points to system corruption. Normally, that error message means one should perform a repair install over the top to fix any components. But new machines? It seems odd — and concerning — that one has to “fix” these so soon after deployment.

To be clear, this is a worry only for a new PC with 25H2.

Once again, Microsoft is being a bit obtuse in its use of AI in its support forums. That post, generated by AI, urges impacted users to “Check for prerequisite updates” when it’s a new machine, in which case any such prerequisites should be handled by the Windows update installer in the first place. The human who steps in to clarify the steps needed to fix the issue does a much better job than AI did to answer the poster’s question.

Always vet AI responses.

Apple

I’ve noted that Apple is starting to push iOS 26, even to older iPhones and tablets. Remember that you can get rid of that glass transparent setting and even have more control over it in 26.2. There is a new slider for controlling the opacity of Liquid Glass throughout the operating system. When you select the “Glass” option, you can use the new slider to make the time look almost entirely clear or give it a more frosted, solid look.

Amazon Alexa

Surprised to see me mention Amazon? Me, too — my Patch Watch work doesn’t usually take me into the world of Amazon devices.

But this is an exception. Not only did I roll back from a forced Alexa software upgrade, I also wasn’t the only one complaining. This latest update caused Alexa’s voice to become too perky. That’s not the end of the world, but it’s like having a person you talked with an hour ago suddenly sound different. It doesn’t happen that often.

But the real problem is that the clock display vanished from my Echo Show devices. What was Amazon thinking? Almost every complaint I saw about the update had to do with the clock!

We have Alexa devices around the house, not only for the clock but also to vocally command groups of lights to be turned on or off. This is supposed to mean I have a smart home, but that will be called into question if Amazon is making stupid blunders like this.

Patch Watch for Dad

Many thanks to all who wished my dad well. He was hospitalized after suffering from low pulse oxygen levels. He’s now back at home with new medications and an oxygen generator — with battery backup — albeit a bit weaker for the experience. Progress, and just in time for Christmas!

I wish you and yours a wonderful holiday season. There’s nothing like family for the holidays!

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 25 December 2025 at 4:27 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

AI coding agents from OpenAI, Anthropic, and Google can now work on software projects for hours at a time, writing complete apps, running tests, and fixing bugs with human supervision. But these tools are not magic and can complicate rather than simplify a software project. Understanding how they work under the hood can help developers know when (and if) to use them, while avoiding common pitfalls.

We’ll start with the basics: At the core of every AI coding agent is a technology called a large language model (LLM), which is a type of neural network trained on vast amounts of text data, including lots of programming code. It’s a pattern-matching machine that uses a prompt to “extract” compressed statistical representations of data it saw during training and provide a plausible continuation of that pattern as an output. In this extraction, an LLM can interpolate across domains and concepts, resulting in some useful logical inferences when done well and confabulation errors when done poorly.

These base models are then further refined through techniques like fine-tuning on curated examples and reinforcement learning from human feedback (RLHF), which shape the model to follow instructions, use tools, and produce more useful outputs.

Over the past few years, AI researchers have been probing LLMs’ deficiencies and finding ways to work around them. One recent innovation was the simulated reasoning model, which generates context (extending the prompt) in the form of reasoning-style text that can help an LLM home in on a more accurate output. Another innovation was an application called an “agent” that links several LLMs together to perform tasks simultaneously and evaluate outputs.

How coding agents are structured

In that sense, each AI coding agent is a program wrapper that works with multiple LLMs. There is typically a “supervising” LLM that interprets tasks (prompts) from the human user and then assigns those tasks to parallel LLMs that can use software tools to execute the instructions. The supervising agent can interrupt tasks below it and evaluate the subtask results to see how a project is going. Anthropic’s engineering documentation describes this pattern as “gather context, take action, verify work, repeat.”

If run locally through a command-line interface (CLI), users give the agents conditional permission to write files on the local machine (code or whatever is needed), run exploratory commands (say, “ls” to list files in a directory), fetch websites (usually using “curl”), download software, or upload files to remote servers. There are lots of possibilities (and potential dangers) with this approach, so it needs to be used carefully.

In contrast, when a user starts a task in the web-based agent like the web versions of Codex and Claude Code, the system provisions a sandboxed cloud container preloaded with the user’s code repository, where Codex can read and edit files, run commands (including test harnesses and linters), and execute code in isolation. Anthropic’s Claude Code uses operating system-level features to create filesystem and network boundaries within which the agent can work more freely.

The context problem

Every LLM has a short-term memory, so to speak, that limits the amount of data it can process before it “forgets” what it’s doing. This is called “context.” Every time you submit a response to the supervising agent, you are amending one gigantic prompt that includes the entire history of the conversation so far (and all the code generated, plus the simulated reasoning tokens the model uses to “think” more about a problem). The AI model then evaluates this prompt and produces an output. It’s a very computationally expensive process that increases quadratically with prompt size because LLMs process every token (chunk of data) against every other token in the prompt.

Anthropic’s engineering team describes context as a finite resource with diminishing returns. Studies have revealed what researchers call “context rot”: As the number of tokens in the context window increases, the model’s ability to accurately recall information decreases. Every new token depletes what the documentation calls an “attention budget.”

This context limit naturally limits the size of a codebase a LLM can process at one time, and if you feed the AI model lots of huge code files (which have to be re-evaluated by the LLM every time you send another response), it can burn up token or usage limits pretty quickly.

Tricks of the trade

To get around these limits, the creators of coding agents use several tricks. For example, AI models are fine-tuned to write code to outsource activities to other software tools. For example, they might write Python scripts to extract data from images or files rather than feeding the whole file through an LLM, which saves tokens and avoids inaccurate results.

Anthropic’s documentation notes that Claude Code also uses this approach to perform complex data analysis over large databases, writing targeted queries and using Bash commands like “head” and “tail” to analyze large volumes of data without ever loading the full data objects into context.

(In a way, these AI agents are guided but semi-autonomous tool-using programs that are a major extension of a concept we first saw in early 2023.)

Another major breakthrough in agents came from dynamic context management. Agents can do this in a few ways that are not fully disclosed in proprietary coding models, but we do know the most important technique they use: context compression.

When a coding LLM nears its context limit, this technique compresses the context history by summarizing it, losing details in the process but shortening the history to key details. Anthropic’s documentation describes this “compaction” as distilling context contents in a high-fidelity manner, preserving key details like architectural decisions and unresolved bugs while discarding redundant tool outputs.

This means the AI coding agents periodically “forget” a large portion of what they are doing every time this compression happens, but unlike older LLM-based systems, they aren’t completely clueless about what has transpired and can rapidly re-orient themselves by reading existing code, written notes left in files, change logs, and so on.

Anthropic’s documentation recommends using CLAUDE.md files to document common bash commands, core files, utility functions, code style guidelines, and testing instructions. AGENTS.md, now a multi-company standard, is another useful way of guiding agent actions in between context refreshes. These files act as external notes that let agents track progress across complex tasks while maintaining critical context that would otherwise be lost.

For tasks requiring extended work, both companies employ multi-agent architectures. According to Anthropic’s research documentation, its system uses an “orchestrator-worker pattern” in which a lead agent coordinates the process while delegating to specialized subagents that operate in parallel. When a user submits a query, the lead agent analyzes it, develops a strategy, and spawns subagents to explore different aspects simultaneously. The subagents act as intelligent filters, returning only relevant information rather than their full context to the lead agent.

The multi-agent approach burns through tokens rapidly. Anthropic’s documentation notes that agents typically use about four times more tokens than chatbot interactions, and multi-agent systems use about 15 times more tokens than chats. For economic viability, these systems require tasks where the value is high enough to justify the increased cost.

Best practices for humans

While using these agents is contentious in some programming circles, if you use one to code a project, knowing good software development practices helps to head off future problems. For example, it’s good to know about version control, making incremental backups, implementing one feature at a time, and testing it before moving on.

What people call “vibe coding”—creating AI-generated code without understanding what it’s doing—is clearly dangerous for production work. Shipping code you didn’t write yourself in a production environment is risky because it could introduce security issues or other bugs or begin gathering technical debt that could snowball over time.

Independent AI researcher Simon Willison recently argued that developers using coding agents still bear responsibility for proving their code works. “Almost anyone can prompt an LLM to generate a thousand-line patch and submit it for code review,” Willison wrote. “That’s no longer valuable. What’s valuable is contributing code that is proven to work.”

In fact, human planning is key. Claude Code’s best practices documentation recommends a specific workflow for complex problems: First, ask the agent to read relevant files and explicitly tell it not to write any code yet, then ask it to make a plan. Without these research and planning steps, the documentation warns, Claude’s outputs tend to jump straight to coding a solution.

Without planning, LLMs sometimes reach for quick solutions to satisfy a momentary objective that might break later if a project were expanded. So having some idea of what makes a good architecture for a modular program that can be expanded over time can help you guide the LLM to craft something more durable.

As mentioned above, these agents aren’t perfect, and some people prefer not to use them at all. A randomized controlled trial published by the nonprofit research organization METR in July 2025 found that experienced open-source developers actually took 19 percent longer to complete tasks when using AI tools, despite believing they were working faster. The study’s authors note several caveats: The developers were highly experienced with their codebases (averaging five years and 1,500 commits), the repositories were large and mature, and the models used (primarily Claude 3.5 and 3.7 Sonnet via Cursor) have since been superseded by more capable versions.

Whether newer models would produce different results remains an open question, but the study suggests that AI coding tools may not always provide universal speed-ups, particularly for developers who already know their codebases well.

Given these potential hazards, coding proof-of-concept demos and internal tools is probably the ideal use of coding agents right now. Since AI models have no actual agency (despite being called agents) and are not people who can be held accountable for mistakes, human oversight is key.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Thursday 25 December 2025 at 4:18 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

On Windows 11, Microsoft has added the ability to manage pretty much all the virtualization features from the new "Virtual Workspaces" page available through the Settings app.

The software giant continues the slow process of porting legacy features to the Settings app, and even though there is still a lot of work to do, the last cumulative update for Windows 11 introduced a new version of the Settings app that includes a "Virtual Workspaces" page.

From this new page, you can enable or disable features like Hyper-V, Windows Sandbox, and more without having to open the legacy "Windows Features" (Turn Windows features on or off) page.

In this how-to guide, I'll outline the steps to get started with the Virtual Workspaces feature on Windows 11.

How to enable virtualization features on Windows 11

To enable virtualization features through Virtual Workspaces, use these steps:

1. Open Settings.
2. Click on System.
3. Click the Advanced page on the right side.
4. Click the Virtual Workspaces setting.

The new hotfix driver is now available for download from the official website under version 591.67. It is based on the most recent driver, 591.59, which was released a few days ago with mostly bug fixes. Apparently, some bugs slid under the radar, and now, NVIDIA is catching up with a quick hotfix release.

Here is the changelog:

This hotfix addresses the following:

 

• Slight banding may be observed on gradients in SDR color mode [5720512]
• Colors are not applied correctly when using Digital Vibrance [5718365]
• Unable to deselect "Show Notification Tray Icon" from NVIDIA Control Panel [5622213]

NVIDIA usually releases new drivers to address bugs that require immediate attention and cannot wait for the next WHQL driver release. However, if you do not have those bugs on your system, you can skip the hotfix altogether and wait for the next driver release. If you want to update from an older driver to the latest one, download the hotfix instead, as it contains everything in the previous driver, 591.59 WHQL.

You can download NVIDIA Hotfix Driver 591.67 from the official NVIDIA support website. Release notes are available on the same page. The driver works on 64-bit Windows 10 and Windows 11 systems with NVIDIA GTX 16 Series and newer. The GTX 10 Series and GTX 900 Series are no longer supported, as NVIDIA recently dropped Pascal and Maxwell graphics card architectures.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 24 December 2025 at 4:04 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

If you have time to play with new operating systems over Christmas and the New Year, you might be interested to know that elementary OS 8.1 is now available. Despite its seemingly minor version number, this is actually a significant update with a focus on hardware support, accessibility, and refined multitasking.

The key change in this update is that Wayland is now the default window manager option and is available under the “Secure Session” selection. Wayland brings several benefits such as better privacy and hardware support. Luckily, you can revert back to the old X11 (“Classic Session”) if you have apps that are incompatible.

This distribution, ever since it was released years ago, has had a strong focus on looking and working well. A smaller new feature in this update now sees password dialogs dim the rest of the screen and prevent focus stealing, this ensures that your entering credentials into a legitimate system prompt.

Elementary OS 8.1 also brings with it the Linux 6.14 kernel and Mesa 25, which improves gaming performance, power consumption, and support for Intel Lunar Lake and M-series Apple Silicon (ARM64).

The AppCenter has received some notable updates in elementary OS 8.1. App pages now show ODRS ratings, controller support for games, and clear labels for in-app purchases. The search in AppCenter has also been improved to be faster and the results are shown in a two-column layout on larger screens. The system updates have also been moved to a System Settings to separate them from app updates.

Aside from these changes, there are lots of improvements to the screen reader and visual accessibility improvements such as improved high-contrast. The dark mode now also includes a snooze feature for timed schedules.

To try out this latest update, head to the elementary OS website. Before being given the download option, you have to choose a donation amount, this can be set to zero if you are not feeling charitable.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Wednesday 24 December 2025 at 4:02 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

This is once again the case this year after releasing a few Insider builds last week. Both the Dev and Beta Channels netted build 26220.7523 for Windows 11, version 25H2 Insiders. It brought a boatload of new features and fixes, including Ask Copilot powered by Microsoft 365 Copilot and Agents on the Taskbar, along with the Agent Launchers framework on Windows. Many of these capabilities are being rolled out in a staggered manner. You can check out the massive change log here, but it's obvious that Microsoft wants AI to become a core component of its numerous Windows experiences.

Prior to that, Windows 11, version 26H1 Insiders on the Canary Channel received build 28020.1362. This was a rare, but equally significant build in terms of new capabilities, as it finally brought the Xbox full screen experience (FSE) to more Windows 11 handheld devices, enhancements for Click To Do, and more configuration in the Settings app, among other things. Check out the full change log for this Canary build here.

With all that said, the Windows Insider team has let users on X (formerly Twitter) know that these were the last flights for this year, which essentially means that we won't be getting new builds until at least 2026. While there is no concrete date known for resumption of Insider builds yet, based on past trends, we expect them to begin quite early in 2026.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Tuesday 23 December 2025 at 3:36 am AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

The Redmond tech giant says that it has always tried to offer the highest level of security with minimal impact to device performance, which is why BitLocker's performance overhead has always been below 10%. However, with NVMe technology getting even better, I/O operations are faster, and consequently, BitLocker struggles to keep up without taking a bigger chunk of CPU cycles. This means that there is a significant performance impact due to the bottleneck that customers notice and obviously don't appreciate.

To work around this problem, Microsoft recently announced hardware-accelerated BitLocker, which will leverage upcoming system on chip (SoC) and CPU capabilities, in addition to maintaining existing support for UFS Inline Crypto Engine technology. There are two new upcoming capabilities that hardware-accelerated BitLocker is focused on:

• Crypto offloading: BitLocker shifts bulk cryptographic operations from the main CPU to a dedicated crypto engine. This capability frees up CPU resources for other tasks and helps improve both performance and battery life.
• Hardware protected keys: BitLocker bulk encryption keys, when necessary SoC support is present, are hardware wrapped, which helps increase security by reducing their exposure to CPU and memory vulnerabilities. This is an addition to the already supported Trusted Platform Module (TPM), which protects intermediate BitLocker keys, putting us on a path to completely eliminate BitLocker keys from the CPU and memory.

Below, you can see how the current implementation (left) differs from the new one (right):

In Microsoft's own benchmarks, the performance differences between the two variations of BitLocker are massive, with hardware-acceleration outperforming software on storage operations like sequential read-writes and random read-writes across the board. There is also a 70% savings in CPU cycles on average when using hardware-accelerated BitLocker as compared to software-based BitLocker, making it almost as good as using your device without BitLocker enabled. The tech firm claims that these saved CPU cycles also result in better battery life for your PC.

You can check out a demo of hardware-accelerated BitLocker below:

As can be seen above, in CrystalDiskMark performance tests, software BitLocker had a read speed of 1632.52MB/s on SEQ1M Q1T1 compared to 3746.55MB/s for hardware BitLocker on the test machine. That's more than a doubling of performance in terms of sequential single-threaded data transfer rate clearly indicating the alleviation of CPU bottlenecking. Write speed also exhibited improvements of 3530.82MB/s compared to 1513.43MB/s. The other random metrics showcased significant speed boosts too.

Microsoft is bringing initial support for hardware-accelerated BitLocker through upcoming Intel vPro devices featuring Intel Core Ultra Series 3 processors, but as the technology matures, it plans to bring these enhancements to all "capable" PCs. This variant of BitLocker leverages the XTS-AES-256 algorithm by default and requires Windows 11 versionb 24H2 or later. However, it will not work in certain other scenarios too, you can check them out here.

Source

Hope you enjoyed this news post. Feedback welcome.

Posted Monday 22 December 2025 at 5:31 pm AEST (my time).

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of November): 5,412

RIP Matrix

Mozilla Firefox is one of the longest-running web browsers, first released in 2004 as an alternative to Internet Explorer, rest in peace. Over the years, Firefox has carved out a sizeable install base, with more than 200 million monthly active users drawn to its speed, deep customisation options, and privacy-focused approach.

Earlier this month, in December 2025, Mozilla appointed a new CEO amid ongoing financial pressure. Unfortunately for long-time fans of the browser, that leadership change comes with a strong push toward AI. Still, there may be a saving grace on the horizon, raising the question of whether Firefox can adapt without losing its identity, or if it is finally time for some users to look elsewhere.

Firefox as an AI browser

It seems like nobody can escape the pull of AI lately, and Mozilla Firefox is no exception. New CEO Anthony Enzor-DeMeo has outlined plans to evolve Firefox into what he describes as a modern AI browser, a direction that has alarmed many users and raised questions about whether it is time to look for an alternative that avoids AI altogether.

Fortunately, at least for now, Mozilla has confirmed that users will be able to disable all AI features within Firefox if they choose. This means AI functionality will be opt-in, rather than something forced on users by default.

That already puts Firefox in a better position than competitors like Microsoft Edge and Google Chrome, neither of which offers a clear, one-click way to disable AI features across the browser fully.

For the time being, Mozilla appears committed to user choice, and being a trusted software company, with the new CEO stating:

Mozilla is dubbing the opt-out of AI features an AI kill-switch internally, giving users a straightforward way to disable any and all AI features within Firefox, with the Firefox Web Developer account on Mastodon stating: