We posted on X a few days back, calling Linux 7.3 one of the most performance-focused releases in years. And now, going through what has actually landed in the merge window since then, that claim does seem to hold up.

The merge window for this release opened right after Linux 7.2 arrived, staying open until the end of this month barring any last-minute happenings that move the closure.

So, here's what is confirmed to go into the 7.3 release so far.

Storage gets a serious speed boost

Btrfs picks up several changes at once. Direct I/O now runs through an IOmap bounce buffer instead of falling back to buffered I/O, taking a workflow that topped out around half of the theoretical maximum throughput up to close to 95 percent of it.

Two separate fixes in the logging and fsync paths add roughly 5x gains each, one from removing an unnecessary delay when multiple logging tasks run at once, the other from skipping unneeded hole detection during full fsync on files with lots of extents but no holes.

The default free-space cache also moves from v1 to v2, the safer and faster option that had not yet become the default.

EXT4 gets its own round of work too. Parallel direct I/O writes are now allowed where the filesystem used to be too restrictive, and the multi-block allocator skips unnecessary work through fallocate().

Underneath both filesystems sits a reworked IOmap layer built around a single iterator callback the compiler can inline instead of dispatching indirectly.

You don't need to do anything for any of this. It just works once you're on 7.3, and the IOmap change pays off the most on fast NVMe, pushing EXT4 from 1.92 million IOPS to 2.19 million in testing.

Important scheduler work

Peter Zijlstra's flatten-the-pick patches rework cgroup scheduling around a single runqueue instead of the old, hitch-prone design. Peter tested it on an aging Sandy Bridge box paired with an AMD Radeon RX 580, and both frame rate and frame-time consistency improved.

If you're gaming on hardware that's seen better days, this is the one improvement that should make it feel less choppy.

Intel hybrid CPUs get fixed too. Cluster-aware scheduling has not handled the P-core and E-core mix well, and Ricardo Neri's work closes that gap on Alder Lake, Lunar Lake, and Panther Lake.

Practically, that means less stutter when your laptop is juggling something heavy like a source build, a game, or a video export, while background tasks compete for the small cores.

Memory management fixes

ZTE engineers found a bug in KSM's reverse-mapping, the step where the kernel checks which memory areas point to a merged page before touching it. Under heavy load, one internal lock could stay stuck for hundreds of milliseconds, freezing other things trying to use it.

The fix drops the worst-case from around 700ms to under 2ms. Simply speaking, it means fewer random freezes on servers and containers under memory pressure.

Xiaomi engineer Wenchao Hao fixed a similar slowdown in zsmalloc, which zRAM and Zswap use to store compressed memory.

A Raspberry Pi 4B saw gains up to 1.83x in testing, a 20-core Intel system up to 1.4x. In plain terms, you get snappier performance on budget boards like the Pi and a smaller but real gain on beefier hardware.

Lower latency for real-time workloads

Every time one CPU core uses the kernel's SMP code to ask the others to run something right away, like flushing memory caches, the kernel used to freeze everything else on that CPU until all the other cores responded.

ByteDance engineers traced the slowdown to that wait and reworked it so other tasks can keep running instead of sitting frozen. On busy systems, that frozen wait could stretch out and stall high-priority tasks.

In fleet testing with DPDK, a framework built for fast networking, the worst-case delay for the slowest requests dropped from ~17ms to ~1.5ms. Most desktop users will not notice this one, but for networking and real-time software, it is an important fix.

Some smaller changes

• RISC-V adds support for eleven newer ISA extensions.
• EFI runtime calls now time out instead of stalling indefinitely.
• RAID 5/6 gets a better benchmark-based algorithm selection at boot.
• New AES library APIs clean up how the kernel handles common encryption modes.

Wrapping up. As you might have read recently, Linus Torvalds spent 18 kernel boots and 24 debug patches chasing a one-line Intel Xe driver bug. AI helped him along the way, and the fix is already merged, of course.

The merge window for Linux 7.3 will stay open for a few more days before the release candidates start rolling out, so there's still room for more to land, and the stable release is expected sometime in October.

David Heinemeier Hansson, aka DHH, has put a number on how serious he is about desktop Linux. That number is eight million dollars, and it's going into a brand new nonprofit called the Omacom Foundation.

DHH is incorporating the foundation now, set to cover the Omarchy trademark, its infrastructure and hosting, promotion for the project, and funding for the open source developers whose work it depends on.

He calls it a "moral obligation," not just to get Omarchy into more people's hands but to change how they use a computer in the first place.

Michael Dell of Dell Technologies, Patrick Collison of Stripe, Tobi Lütke of Shopify, Jack Dorsey of Block, Matthew Prince of Cloudflare, Jason Fried of 37signals, Brendan Iribe of Sesame, and DHH himself each put in a million dollars to make it happen.

People already like using Omarchy, DHH says, and he wants this money to last as long as possible instead of getting spent all at once while also making the most out of it.

Hyprland first to gain

Hyprland is the first project outside Omarchy to see any of this money. Omacom is extending an exclusive sponsorship to support Vaxry, the developer behind the Wayland compositor Omarchy runs on.

Under this, he is locked into an agreement for three years, with the possibility of extending for two more years.

This way Vaxry gets to work on Hyprland full time without needing to think about the money. As part of this arrangement, the €5/month Hyprperks subscription service will be discontinued, and everything offered under it will be provided for free.

DHH also joked about finding Vaxry a pookie, noting that:

Many thanks to 37signals (hello self! 😂), Butterfly, and Framework for helping Hyprland get to this point. Thrilled that the Omacom Foundation can take over from October 10, and give one of the best developers in this new Linux moment the chance to go all-in.

(* This deal does not include any direct assistance in getting Vaxry a girlfriend, but I’ll personally be available as a reference for any inquiries on the matter.)

Do you remember this?

This isn't the first time we are talking about money and Hyprland. Back in 2025, Framework became a Gold tier sponsor of the project. But that didn't sit too well with many people, as soon after, a thread titled "Framework supporting far-right racists?" came up on the official forum.

The original poster called out Vaxry's old freedesktop.org ban and pushed Framework to explain why it was funding the project. Framework's own account made things worse that same day. It posted a tweet comparing Omarchy to the old Windows XP product key meme, and people read that as an endorsement of DHH.

While Framework did send hardware to people at Omarchy, the project itself never got any direct funding. Later, Framework had to start publishing a running list of its sponsorships after people kept asking exactly who the company was funding and why.

Where Omarchy's headed

Quattro, Omarchy's newest release, shipped recently, building AI agents into the desktop by default instead of treating them as an extra. While none of the eight patrons have said publicly that Omarchy's AI bet factored into their decision to back the foundation.

My read is that it's hard to ignore, especially with Michael Dell's company betting heavily on AI infrastructure elsewhere. A distro built around AI agents is exactly the kind of bet a company selling AI hardware would want to back.

And it all ties into what DHH said about truly democratizing Linux.

Whether that $8 million and a nonprofit structure are what bring about the "year of the Linux desktop" is a question for another time.

Suggested Read 📖: Speaking of funding, Canonical has pitched in for a 3-year PhD project that looks to automate C to Rust translation for important Ubuntu components.

Built around servers people can host, TeamSpeak has been around since 2001, serving as a reliable communication tool that has gained a reputation for running light and reliably over inconsistent networks.

It has gone through many iterations that have seen it grow from a simple client-server voice tool into something that has a modern interface, built-in screen sharing, and the option to buy hosted servers.

TeamSpeak 6, currently in beta, is the most significant overhaul the platform has ever pushed, having entered testing in January 2025 and progressed through several builds to today's 6.0.0-beta4.1.

If you assume that the platform only focuses on mainstream OSes, then you would be wrong. The client and server software have supported Linux alongside Windows and macOS since the early versions.

Now the developers have announced that direct calls are coming.

Direct calls will be possible

TeamSpeak is adding the ability to make 1-1 direct calls, letting two people call each other without setting up or joining a server first. It's not live yet since the devs are still testing it, but they say it's coming soon.

That's not the only thing direct calls are coming with. Screensharing is built right in, letting you share your screen, a game, or your camera straight from the call at up to 4K or full source quality.

Presets facilitate a quick setup, or you can select the resolution, bitrate, FPS, and audio if you'd rather configure it yourself. There's also a privacy setting so you can choose whether the stream is public, limited to contacts, or private, plus a toggle for capturing system audio alongside it.

Picture-in-picture comes with it too. Using the Shift key, you can drag up to four screenshares or cameras into a corner of another feed while sharing, arranging them according to your preference.

When asked about the possibility of making group calls, SYOX, the Community Manager at TeamSpeak, said that:

This feature focuses on 1-1 direct calls between two users in a private chat. It remains to be seen whether and to what extent we will introduce Group Calls. We are definitely keeping an eye on group calls as a feature addition, but the scope has not yet been discussed, as it was not the focus of this release/implementation.

Can I finally move away from Discord?

This change has me wanting to try TeamSpeak again. Direct calls without needing a server is precisely the thing that made my last test of it so brief, since spinning up a server just to talk to a person never felt worth the effort.

If the feature works as well as it looks in these previews, I might end up ditching Discord sooner than later. And if they can figure out group calls in the coming months, then I won't even need to host a server to talk to my friends, so that's a win too!

The GNOME Extensions ecosystem is packed with tools that let you customize almost every part of the desktop. Whether it's changing the appearance of the top panel, replacing the app launcher, adding visual effects, or making the interface behave exactly the way you want.

We have covered extensions in the past:

• We created a generic list of best GNOME extensions to show extensions of all kinds
• And then we also have a list of GNOME extensions for productivity

Since there are so many of them, I create this new list that focuses on extensions for desktop clooks customization.

1. Just Perfection

If there is one customization extension that deserves to be installed first on a fresh GNOME setup, it's Just Perfection.

As the name suggests, the extension lets you tweak countless parts of the GNOME Shell. From small visual adjustments to larger interface changes, it gives you an impressive amount of control without requiring you to edit configuration files.

In fact, the number of available tweaks is so extensive that it deserves an article of its own, which we already have and you can read it here.

The screenshot below contains effects of another GNOME Extension called User Accent Color which we will see in a later section.

Some of the highlights include:

• Change the position of the panel, clock, and notification indicators.
• Adjust the size and corner radius of the Activities Overview.
• Show or hide individual panel items.

And that's just a glimpse. I recommend checking the project's official README to explore its exhaustive list of features.

2. User Themes

Installing a GTK theme changes the appearance of your applications, but it doesn't affect the GNOME Shell itself.

That's because the Shell uses its own theme, separate from GTK. If you want to customize the top panel, Activities Overview, Quick Settings, and other Shell components, you'll need the User Themes extension.

Once installed, you can apply any compatible gnome-shell theme stored in your ~/.themes directory using the GNOME Tweaks application.

It's a small extension, but an essential one if you're interested in desktop theming.

3. Wallpaper Slideshow

If you like refreshing your desktop without manually changing wallpapers every few days, Wallpaper Slideshow is an excellent extension to have.

It automatically cycles through images from a folder of your choice at configurable intervals, turning your wallpaper collection into a dynamic slideshow. The extension also supports downloading Bing's Wallpaper of the Day if you prefer fresh backgrounds without maintaining your own collection.

Better yet, it automatically detects changes to your wallpaper folder, so newly added images become part of the rotation without any extra setup.

You can control the slideshow from the Quick Settings menu or the desktop context menu, and the extension even remembers its queue across reboots, allowing the slideshow to resume where it left off.

It's a simple way to keep your GNOME desktop feeling fresh with minimal effort.

4. User Accent Colors

One of the nicest visual touches in modern desktop environments is having the system's accent color automatically match your wallpaper. If you've wanted the same experience on GNOME, User Accent Colors is worth a look.

The extension extracts the dominant accent color from your current wallpaper and applies it across the desktop, including GNOME Shell, Adwaita applications, and adw-gtk3 themed GTK3 applications.

It can also generate matching tinted themes and even recolor supported icon packs, giving the entire desktop a much more cohesive appearance.

The result is a desktop that automatically adapts whenever you change your wallpaper, without manually tweaking themes or accent colors. If you're building a visually consistent GNOME setup, this extension is an excellent finishing touch.

5. Dash to Dock

By default, GNOME only displays the Dash while you're in the Activities Overview.

Many users, however, prefer having a dock that's always visible and keeps their favorite applications within easy reach.

Dash to Dock transforms the built-in GNOME Dash into a fully customizable dock that can be positioned on any edge of your screen. You can tweak its size, behavior, transparency, auto-hide options, and much more.

In fact, Ubuntu's default dock is based on a customized version of Dash to Dock.

The only thing I personally miss is a ripple-style hover animation. Fortunately, the next extension offers a different take on desktop navigation.

6. Dash to Panel

While Dash to Dock focuses on creating a dock, Dash to Panel takes things a step further.

It merges the GNOME top bar and Dash into a single panel, giving the desktop a layout that feels much closer to Windows.

If you're switching from Windows 10 or Windows 11, this extension makes GNOME feel instantly familiar.

Beyond changing the layout, Dash to Panel offers a surprising amount of customization. You can:

• Change the panel's color, opacity, and appearance.
• Show or hide individual panel items.
• Keep the original GNOME top panel while using only the dock functionality.
• Customize various behaviors to suit your workflow.

It's one of the most feature-rich panel extensions available for GNOME.

7. Dhruva Dock

Dhruva Dock is the newer addition to the GNOME extensions ecosystem, and it has quickly gained attention thanks to its polished appearance and smooth animations.

The dock supports true magnification effects, includes an integrated clock, and offers several preset themes alongside color customization options.

Unlike many dock extensions, Dhruva Dock isn't limited to the screen edges. You can position it almost anywhere on your desktop, giving you much more flexibility when designing your workspace.

Its fluid animations are easily one of its strongest selling points.

8. Blur My Shell

Few visual tweaks change the look of GNOME as much as adding blur effects, and Blur My Shell has become the go-to extension for that purpose.

What started as a simple blur utility has gradually evolved into a comprehensive customization tool. Besides adding blur to the Activities Overview and various Shell components, recent versions even allow blur effects to be applied to application windows.

The extension also lets you configure different blur pipelines, giving you fine-grained control over which parts of the desktop should be blurred and by how much.

While it doesn't offer the level of control found in Wayland compositors like Hyprland, and older hardware may struggle with the added effects, it's still one of the best visual customization extensions available for GNOME.

9. Dynamic Music Pill

If you frequently listen to music while working, Dynamic Music Pill is an extension worth checking out.

Instead of displaying a simple scrolling track title in the panel, it presents your currently playing music inside an elegant pill-shaped widget with playback controls.

You can place it either on the top panel or alongside the Dash, depending on your preferred layout.

Beyond its attractive design, the extension also includes several useful features:

• Adaptive colors based on the album artwork.
• Multiple visualizer modes.
• Playback controls with configurable click actions.
• Synchronized lyrics.
• Transparency and appearance settings.

It's a great combination of functionality and visual polish.

10. Top Bar Organizer

If you've ever wished you could rearrange the items on GNOME's top panel, Top Bar Organizer makes that incredibly easy.

Unlike many customization extensions, it doesn't overwhelm you with dozens of configuration options. The interface is refreshingly simple.

You simply drag and drop panel items between the three available sections:

• Left Top Bar Box
• Center Top Bar Box
• Right Top Bar Box

That's all there is to it.

Whether you want the clock on the left, system indicators in the center, or a completely custom layout, this extension makes it possible in just a few clicks.

11. Arc Menu

GNOME's built-in search is one of its strongest features. It lets you quickly launch applications, find files, perform calculations, and much more, all from a single interface.

That said, many users still prefer having a traditional application launcher that's always available.

Arc Menu is much more than a simple Start Menu replacement. It offers an impressive collection of menu layouts inspired by different desktop environments and operating systems.

Want a Windows 11-style Start Menu? That's available. Prefer a macOS-inspired launcher? You can switch to that with just a few clicks.

Beyond the layouts, Arc Menu provides extensive customization options. You can change icons, tweak the appearance, reposition the menu, customize categories, and fine-tune almost every aspect of the launcher.

If you enjoy experimenting with different desktop layouts, Arc Menu is one of the most versatile extensions you can install.

12. Burn My Windows

Even the developers describe Burn My Windows as a "useless" extension.

I couldn't disagree more.

If you enjoy desktop animations, this extension is pure fun. Instead of the standard open, close, and minimize animations, it lets you replace them with dozens of eye-catching effects.

Some of my favorites include Apparition, Broken Glass, Incinerate, and Matrix, though there are plenty more to experiment with.

It won't make you more productive, but it definitely adds some personality to your desktop.

13. Customize Clock on Lock Screen

The GNOME lock screen doesn't offer much room for customization out of the box.

This extension changes that by letting you customize the appearance of the lock screen clock.

One particularly interesting feature is the ability to display the output of shell commands. That opens up several possibilities, such as showing inspirational quotes, system information, weather updates, or any other text generated by a script.

For example, I used the command below to print the system details as shown in the screenshot above:

printf "🧠 CPU: %s | 💾 RAM: %s | 💿 Disk: %s" "$(top -bn1 | grep 'Cpu(s)' | awk '{print $2 + $4 "%"}')" "$(free -h | awk '/Mem:/ {print $3 "/" $2}')" "$(df -h / | awk 'NR==2 {print $3 "/" $2 " (" $5 ")"}')"

If you enjoy personalizing every part of your desktop, this extension gives the lock screen a lot more flexibility.

14. Compiz Windows Effect

If you used Linux during the Compiz era, chances are you still remember the famous wobbly windows.

This extension brings that classic effect to modern GNOME.

Whenever you move or resize a window, it behaves like a flexible sheet rather than a rigid rectangle, recreating one of the most iconic desktop effects from the early days of Linux desktop customization.

It certainly makes interacting with windows a lot more enjoyable.

15. Compiz Magic Lamp Effect

Continuing the nostalgia, Compiz Magic Lamp Effect recreates another classic animation from the Compiz days.

Instead of windows simply disappearing when minimized, they smoothly stretch and collapse into the dock using the familiar Magic Lamp animation.

The effect looks particularly good when paired with dock extensions like Dash to Dock, making window animations feel much more polished than the default GNOME behavior.

If you're already using one of those docks, this extension is a perfect companion.

16. Background Logo

Sometimes the smallest tweaks can make your desktop feel a little more personal.

Background Logo lets you place a logo or image directly on top of your desktop wallpaper without modifying the wallpaper itself. Whether you want to display your favorite Linux distribution's logo, your organization's branding, or a custom SVG, the extension makes it easy.

It may not be an essential customization, but it's a nice option for users who enjoy putting their own stamp on every part of their desktop.

17. Modern Clock

If you like the clean desktop widgets found in KDE Plasma, Modern Clock brings a similar experience to GNOME.

Instead of placing the time in the top panel, this extension adds a modern-looking desktop clock widget inspired by KDE's Modern Clock. It sits directly on your desktop and blends nicely with minimalist wallpapers, making it a subtle yet attractive addition to your workspace.

Unlike the standard GNOME clock, this one is meant to be part of your desktop rather than the shell interface. If you're building a customized desktop with widgets and visual enhancements, Modern Clock fits right in.

Wrapping Up

One of GNOME's greatest strengths is the flexibility offered by its extension ecosystem. While the desktop intentionally keeps its default experience clean and minimal, these extensions allow you to shape it into almost anything you want.

Whether you're looking for a Windows-like taskbar, a macOS-inspired launcher, beautiful blur effects, playful window animations, or simply more control over the interface, there's an extension here that can help.

Of course, you don't have to install all of them. In fact, I'd recommend starting with a few that solve problems in your daily workflow and then gradually adding more as needed.

Did I miss one of your favorite customization extensions? Let me know in the comments. I'd love to check it out.

Linus Torvalds just used AI to help fix a Linux kernel bug.

Let me tell you that this is a rare event. Not because Torvalds is using AI. Last Christmas holidays, Torvalds used AI for the first time for a personal, hobby project. He was averted to AI at first but made his peace with it, even for AI-assisted code contributions in the kernel project.

The event we are talking about is rare because Torvalds doesn't code on Linux kernel anymore. He is in the maintainer mode. He just reviews code and merges pull requests from various contributors to the kernel code.

So this is a rare occasion in recent time when he made a kernel patch himself. And he did that with the assistance of AI.

Fixing a one-liner bug with AI

Don't scoff at "one-line" code change. Speak to any software developer and they will telly you that often the bugs are easy, few line code changes. It's the process of finding the root cause that takes time, effort, and skills.

The bug Torvalds dealt with was in the Intel Xe graphics driver for Battlemage G21 cards. It caused GDM, the display manager, to restart endlessly.

Tracking it down with AI required 24 debug patches (did he just use "printfs?) and 18 kernel boots. Ultimately, the culprit turned out to be a single line where round_up() should have been round_down().

And this was a debug session from hell, enormously helped by an AI doing much of the grunt-work.

The AI "did the grunt work" for Torvalds by adding debug instrumentation, running analysis, and even writing the final commit message. I can feel the commit message part, it's a boring but necessary task.

But this debugging session was not all smooth sailing. Torvalds didn't just add "don't make mistakes" to his prompts. On the contrary, the AI told Linus multiple times that the problem was "impossible and unsolvable." Torvalds refused.

I suspect those things have been trained by people who may not be quite as stubborn as I am.

AI is a tool, use it effectively

This is what many people in the software industry understand. AI is just a tool and it is essentially useful when there is a capable human driving the process.

Look at the example here with Torvalds. A vibe coder with no knowledge of what's going on with the code, would have left it to AI to make decisions, and AI would have called the bug unfixable and walked away.

But Linus Torvalds has experience. He could understand what the AI was doing right or wrong. He persisted and fixed the bug.

The patch will be in Linux 7.3 and has been marked to be backported to stable kernel branches.

The creator of Linux needed 18 boots to find a one-line fix. The AI helped him get there. Although, I think he would have gotten there all by himself as well, even if it would have taken a bit longer.

There is no information available on what model or tool he used. In his personal project, where he used AI for the first time, he had used Gemini from Google. I guess he is not into local AI for now. I hope he finds enough interest and gives us a new gift after Linux kernel and Git. Wishful thinking, I know.

Source: Phoronix.

Proton Mail (partner link), the service people turn to for escaping Big Tech, has rolled out "Categories" as a new inbox view that automatically sorts incoming email into six groups: Primary, Social, Promotions, Newsletters, Transactions, and Updates.

When you receive a new mail, it gets sorted automatically into one of six categories based on metadata, without Proton reading the contents of the message. This is what Gmail already introduced years ago and it does work for most parts.

How does it work?

By default, Primary houses personal and work-related messages along with any important updates; Social covers social media updates and activity notifications; Promotions covers discounts and sales offers; and Newsletters carries news content you signed up for.

The other two categories can be enabled via the Settings if you require them.

Of course, we are not claiming that this is something revolutionary. Gmail and Outlook have offered inbox tabs for years, but they achieved that by scanning the contents of emails that are hosted on their servers.

Proton's approach here is different, sorting mail by metadata instead of reading what's inside each message, since its zero-access encryption keeps that content unreadable even to itself.

Anant Vijay Singh, the Product Lead for Proton Mail, goes even further to note that:

Proton’s business model aligns our incentives with yours because we’re 100% funded by our community, not by advertisers. Enabling Categories doesn’t change any of this.

Your inbox remains protected by zero-access encryption, and emails you send with other Proton users (or using password protection) are end-to-end encrypted.

It's already live

If you haven't logged into your Proton Mail (partner link) account recently, then keep an eye out for a dialog titled "Your inbox, automatically organized." It will show you two options; go with "Yes, organize it" if you want to upgrade to the new category-focused inbox experience.

If you'd rather stay on the legacy, unified inbox system, then choose "Keep inbox as before."

Once enabled, you will see the categories all lined up at the top of your emails, with a useful blue dot appearing on the ones with unread messages.

Moving an email to a different category takes a right-click followed by "Move to category..." or a drag and drop. Going forward, Proton Mail sorts similar emails into that category automatically.

If the default category selection or the feature itself doesn't suit your preferences, then you can manage it from the Settings menu (gear button), where you could either disable it by clicking on the button near "Email categories."

Or, you could go into the "Email categories" settings page to enable or disable individual categories and toggle notifications for each one. I was content with the default selections, so I didn't bother changing anything.

You can see how Proton Mail categorized my mail. 👇

Keep in mind that you cannot disable the "Primary" category as that one always stays on, catching anything that needs your attention along with emails from any category you've turned off.

And that's not all. Proton is also bringing full-content search to Proton Mail on mobile in the coming weeks. It'll let users search the actual contents of their emails, not just subjects and senders.

Proton says that the search will run entirely on the user's device rather than touch its own servers.

Canonical has committed funding to a three-year PhD project focused on building a system that can automatically translate large C codebases into Rust.

And they are not alone; UK Research and Innovation is matching their funding for the project, which is set to run through the University of Bristol's Programming Languages Research Group.

The PhD is aiming to build an all-encompassing platform that can take a C repository running into hundreds of thousands of lines and translate it into Rust that's "safe, behaviourally correct and maintainable Rust."

The work will be carried out by a student, Alex Wood, who will be supported by Professor Meng Wang leading the work, with Dr. Cristina David and Canonical's Jon Seager serving as co-supervisors.

How would this work?

Getting from C to Rust automatically means solving a few separate problems, and the system tackles them in sequence. First, it has to break a massive codebase into pieces small enough to work with without losing track of how those pieces depend on each other elsewhere in the code.

Only then does the actual translation happen, handled by language models that have already been trained on real C-to-Rust conversions, with the goal of writing code that looks like something a Rust developer chose to write rather than C dressed up in Rust syntax.

That's the easier half. The harder task is proving the translated code actually behaves like the original, which is where fuzzing and formal equivalence checks come in.

When something doesn't line up, a separate repair step tries to track down exactly where the translation went wrong and patch it using symbolic program-repair techniques instead of starting the whole chunk over.

Jon points out that the language model handles only one part of the job. Whatever it generates still has to clear validation before anyone can trust that it kept the original behavior intact.

The targeted test case

Both AppArmor and snap-confine already run in production on real Ubuntu systems, handling access control and sandboxing that most users never see but rely on constantly.

That's why they were picked, not because they are simple, but because Canonical wants industrial-grade test cases that offer a much better look at how the approach holds up under real production conditions.

Testing a live component carries with it everything that's built up around it through years of real use, which makes it a far harder thing to translate without breaking something, and that's the point.

Jon also flagged an important caveat, noting that nothing generated by this project is going into either tool anytime soon. The point, he said, is that Canonical has "a vested interest in the software and are keen to see the results," not that a rewrite is coming.

Why should you care?

If the approach survives this test, Canonical will have a real answer to whether automated translation can be trusted with code that protects the system, not just the code that was easy to test on.

As a result, Ubuntu users, from regular desktop installs to enterprise deployments, could eventually see security-critical components with fewer bugs without Canonical having to gamble on a time-consuming manual rewrite of sensitive code.

LightDM 1.33.0 is out after a four-year gap, now under community maintainership rather than Canonical's. Two new maintainers, one behind Ubuntu Cinnamon Remix and the other a long-time Fedora and openSUSE contributor, have stepped up.

Nitrux has given a name to what it's been building since dropping Plasma for Hyprland last year. The Workspace Environment brings four in-house QML components together, a workspace bar, a dock, a settings app, and a lock screen.

A busy point release for Linux has just landed. AMD Radeon picks up HDMI 2.1 support, Apple M3 MacBooks can boot Linux for the first time, and Btrfs large folio support is on by default.

Canonical's VP of Engineering has said that Ubuntu's growth rate inside WSL is outpacing its growth rate on native desktops, which got picked up everywhere (including here 🤐) as "Ubuntu is growing faster on Windows than Linux." He's since clarified that the comparison was about growth rate only.

The Xen Project recently launched a Safety Committee to build the shared documentation, test evidence, and certification artifacts that companies deploying Xen in automotive, avionics, and industrial automation keep having to produce independently.

Privacy company Murena now offers Fairphone Gen 6 with its /e/OS, a private alternative to Android.

PINE64 is halting the production of its Linux-based tablets. Reason? The soaring RAM prices have made it impossible to keep the pricing reasonable for such hobbyist devices. And the RAM prices are increasing due to AI boom. Big Tech has already put their hands on the supply with bulk orders for their data centers and now the consumers are struggling as prices for disks and RAMs have gone up multiple folds.

🧠 What We’re Thinking About

Don't remember if I shared it in the newsletter or not. A few days ago, it was found that some AI companies are buying rare books to train their LLMs. And in the process, they were also cutting (read destroying) the books because it is easier to scan the books page by page that way.

404 Media (paywalled content) has tracked one such shipment to the Amazon AI training facility. So at least we have one culprit identified.

🧮 Linux Tips, Tutorials, and Learnings

Most people look no further than Tmux and Screen. But what if I told you there were terminal multiplexers that shared sessions over SSH, restored your workspace exactly as you left it, or brought tiling WM concepts straight into your terminal?

A detailed guide on apt that covers updating the package database, installing and removing packages, searching, showing package info, listing upgradable software, and cleaning up leftover dependencies.

Want your GNOME desktop to look like those Catppuccin or Nord riced screenshots without spending an afternoon in config files? Rewaita is a Flatpak app that applies color palettes to GTK and Libadwaita apps through CSS overrides.

👷 AI, Homelab and Hardware Corner

Proton has come up with a tool that tells you how many secrets you have spilled over to an AI chatbot.

✨ Apps and Projects Highlights

stillOS is a new distro that's built on an AlmaLinux base with atomicity as its main selling point.

📽️ Videos for You

I show you a wonderful dock for GNOME users. See it in action in this week's video.

💡 Quick Handy Tip

On GNOME inside the Nautilus file manager (aka Files), press CTRL+F to start a search in the current directory and CTRL+SHIFT+F to search everywhere. You can also select the result and use CTRL+ALT+O to open that location in file manager.

🎋 Fun in the FOSSverse

If you went through the apt guide earlier, why not test your knowledge with a quick quiz?

The beginning is always simple. 😆

🗓️ Tech Trivia: On August 22, 1955, representatives from seventeen companies that had bought IBM's new 704 mainframe met at RAND Corp in Santa Monica and formed SHARE, the first computer user group, so members could trade code and documentation instead of duplicating each other's work.

🧑‍🤝‍🧑 From the Community: A long-time FOSSer has returned, kicking off a thread on how Windows enabled bitlocker without asking him, resulting in considerable wear on the attached storage disks.

The open hardware manufacturer recently posted an update on its Mastodon account, and it's not great news if you were hoping to grab new Linux-powered PINE64 hardware.

The disruptions in DRAM and eMMC supply are to blame, as PINE64 says its Linux hardware production pipeline is empty for now, with nothing new planned while component costs stay high.

Whether that changes depends on component pricing after mid-2027, and they aren't promising anything before then. They expect stocks for the PineNote and PineTab2 to run out in three months.

Once that happens, there's no more new launches or restocks to cover the outgoing stock.

Not everything is affected though. PINE64's microcontroller lineup, the PineTime, PineVoice, and Pinecil, will keep shipping as usual since they don't rely on the same memory chips.

Concluding their announcement, they gave everyone a status update on the upcoming PineTime Pro:

As for the PineTime Pro, currently the demo firmware has basic drivers for the display, touch, motion, GPS, buttons, heat rate sensor (chip id only), battery charger, motor, microphone and the speaker. Though some issues do need to be ironed out first before a pilot batch begins production.

That's it for this post, all in all it's a bit bleak for the Linux devices, but in the meantime PineStore still have a great lineup of MCU powered devices which will continue to expand with more exciting devices.

The unnamed culprit

PINE64 never actually says the word "AI" in their post, choosing to keep the announcement focused on the DRAM/eMMC shortage. But you have to look further to see where that shortage is actually originating from.

Memory prices have been climbing for months, and DDR5 has taken the hardest hit of all. Companies running AI data centers have been locking up global memory supply well into next year, leaving everyone else fighting over what's left.

Executives across the memory industry aren't optimistic either. SK Hynix's CEO recently said that the worst is still ahead, with supply expected to lag behind demand for years to come.

Not the only ones

PINE64 isn't the only Linux-focused hardware provider caught in this price surge.

The Orange Pi Neo, a Linux-first handheld built with Manjaro, has been sitting on ice since earlier this year for the same reason. The project has been in development for two years and still has no launch date, all because DDR5 and SSD prices refuse to come back down.

I am sure these aren't the only projects affected; smaller ones still in the crowdfunding or conceptualization phase must be facing the same problem.

What worries me is that the AI bubble will eventually burst, but small Linux-focused projects are the ones getting choked out right now. If these component prices stay this absurd, we are going to lose the few independent hardware choices we actually have left.

Suggested Read 📖: Omarchy Bets Its Future on AI Agents While the Linux World Stays Cautious

How software gets built at an organization has evolved a lot over the years. Long before AI agents entered the scene, engineering teams were already turning development into a series of repeatable actions that brought together shared components and defined processes while minimizing reinvention.

That approach has made teams faster and more consistent. In the age of AI, it needs to transform yet again. Warp is convinced it has figured out what's next.

The company has introduced Warp Factories as an open infrastructure offering meant to help engineering teams build and run their own cloud software factories. Here, fleets of coding agents do the work instead of an agent limited to a developer's laptop.

The short explanation is that it's Warp's attempt to turn "run an agent in the cloud" into an actual assembly line, with clear steps, real numbers to track, and humans stepping in only when it matters.

An assembly line for development

Warp Factories acts as an infrastructure layer for running fleets of coding agents across the entire software development lifecycle. In Warp's own framing, that makes it a cloud software factory, an automated loop where agents handle the work from intake through to a shipped result.

Agents handle triage, write specs, implement changes, review the work, verify it, and monitor the result, while humans only step in at decision points that actually need a person to verify what's been done.

Think of it like a car assembly line. One station welds, another paints, another does final inspection, and the vehicle doesn't move forward until the station before it signs off. Warp's version applies that same logic to code.

One agent triages an incoming ticket, another drafts a spec if the change is complex enough to need one, another implements the fix, and another reviews it before a human approves the merge.

Warp is leaning hard on the idea of AI sovereignty to explain why this needs to exist as infrastructure. Teams can bring their own hosting, and every agent conversation, evaluation, and memory record can be kept on infrastructure the company controls instead of Warp's.

Data can also be kept out of any training pipeline entirely through zero-data-retention options.

The founder and CEO of Warp, Zach Lloyd, further points out that:

Our vision is providing you all with the core components to get to this point as quickly as possible. In an age where time to ship is arguably your most important competitive advantage, you should be standing up and operating a factory today, not six months from now.

More of its capabilities

Factories also ships with a dashboard the company calls the "factory app," basically a control room where you can watch live and past agent runs, check what's pending, and review every automation in one place.

You're not locked into Warp's own models or its own agent either. Factories works with any model, open-weight ones too, and any harness, so Claude Code or Codex can sit in the implementation seat instead of Warp's native agent.

All of that activity gets measured automatically, with cost, throughput, and code quality metrics feeding into evaluation and benchmark tools that track how a factory's performance shifts over time.

Get started

Warp Factories isn't available for everyone right now. If you or your organization wants to try it out, then you will need to request access through Warp Factories' early access portal.

Qualifying organizations can get up to $10,000 in factory usage credits.

We are used to seeing Murena roll out new devices every now and then. Their whole sales pitch revolves around a de-Googled, privacy-focused Android experience. Not everyone is sold on that approach, sure, but the company has managed to stick around for quite a few years now, which says something.

The latest addition to their lineup is the Fairphone (Gen. 6+) with /e/OS. It sits right next to the standard Gen. 6 model, which is still on sale. The new variant packs better hardware and carries a higher price tag to match.

The new device

The most significant change is the Snapdragon 7s Gen 4 processor, paired with 12GB of RAM. The base model makes do with the older Gen 3 chip and 8GB of RAM. Though storage stays at 256GB across both, expandable via microSD up to 2TB.

Everything else on the hardware carries over from the base model, including the 6.31-inch 120Hz LTPO OLED display, the 50MP main camera with a 13MP secondary and 32MP selfie shooter, the 4,415mAh removable battery with 30W charging, IP55 water resistance, and Gorilla Glass 7i to protect the display.

There's also a new Cobalt Blue colorway that ties into Fairphone's fair-mined cobalt supply chain efforts. It joins the existing Forest Green and Horizon Black options. The modular design is unchanged, so you can still swap out 12 key components with a regular screwdriver.

The software itself has not changed for this launch. It is still the same privacy-first build Murena has been refining for a while now.

That means no Google tracking baked into the OS, an Advanced Privacy widget that lets you control what data individual apps can see, and a running Wall of Shame that calls out installed apps leaking the most personal information.

The founder and CEO of Murena, Gaël Duval, spoke about the device, saying that:

With devices like Fairphone (Gen. 6+) with /e/OS, users can have a truly ethical device both in terms of hardware and software.

When I developed Mandriva Linux back in the 90s, it was my mission to make the operating system usable for the everyday, non-techie user. This is my goal with Murena – to give the average user access to smartphones that do not hold their privacy to ransom.

Get yours

Pricing depends on which region you order a Murena Fairphone (Gen. 6+) from. For EU, it costs €699, more than the €649 Murena charges for the standard Gen. 6. In the US, the Fairphone (Gen. 6+) is priced at $699, which is more than the standard Gen. 6's $649 price tag.

People who order during the limited-time launch window of 18 August-1 September 2026 get a few extras. In the EU, that's two free accessories of your choice, where you get to pick from a case, screen protector, finger loop, card holder, or lanyard. US buyers get a case and screen protector bundle.

Everyone gets a year of 1TB Murena Workspace and a year of Murena MDM.

Availability differs by region too. The device is shipping now in Europe, with the US listed as a backorder.

Closing thoughts

The ethical smartphone market is still tiny, and it mostly survives on partnerships like this one. Fairphone builds the hardware, Murena handles the software, and together they keep proving that a privacy-first phone does not have to mean a phone stuck in the dark ages.

Whether that is enough to pull people away from Google remains the same open question it was the last time Murena updated their lineup.

Atomic distros work differently from the Fedora or Ubuntu install you're probably used to. Instead of layering individual packages on top of each other, the entire base system ships as one image, and updates land as a whole swap rather than a slow stream of package changes.

As a result, you get a system that's hard to break, where updates apply cleanly or not at all, rolling back a bad patch is easy, and there's no dependency hell to wade through since every install starts from the same known-good state.

The trade-off is flexibility. You lose the ability to tinker with individual system packages the way you can on a traditional distro, and anything outside the base image needs a workaround instead of a quick apt install or dnf install.

stillOS wants you to consider the upside of that trade-off. It calls itself a "maintenance-free" Linux desktop, built on an AlmaLinux 10 base with bootc handling the atomic updates.

Behind it is Cameron Knauff, who goes by the name "PizzaLovingNerd" in the Linux community, and this isn't his first distro.

His earlier project, risiOS, a Fedora-based distro aimed at smoothing out the rough edges for new users, picked up a solid following before it was absorbed into Nobara Linux, freeing him up to start stillHQ and, down the line, stillOS.

It has potential

I tested stillOS 10.2 on a virtual machine in GNOME Boxes, running on my Fedora Workstation setup. Installation felt familiar right away. stillOS uses the same Anaconda installer AlmaLinux ships with.

If you've installed Fedora, RHEL, or any of their downstream distros before, you already know the steps involved here.

Onboarding is quick and doesn't waste your time. A short welcome screen kicks off Quick Setup, which walks you through installing apps, picking a browser, and grabbing essentials before dropping you onto the desktop.

Once you boot into stillOS, you will see that while the desktop is GNOME underneath, the distro ditches the usual top-panel layout for a taskbar-first interface, closer to Windows 11 than anything GNOME ships by default.

The taskbar carries the usual options like the app drawer, pill-shaped workspace switcher, an update notifier, pinned apps, the calendar, the clock, and quick settings.

Unlike vanilla GNOME, stillOS' version lets you drop icons straight onto the desktop.

Right-clicking gives you quick access to creating new folders and documents, undoing an action, selecting everything, arranging icons, opening the terminal, and some important settings.

The file manager keeps most of that same right-click menu, folders, documents, select all, opening a terminal, alongside the usual "Properties" entry found on Nautilus (Files).

The document creation option is the one that I find very useful, as that’s something I miss on the vanilla GNOME experience. The catch is that it only hands you three formats through ONLYOFFICE: docx, pptx, and xlsx. Considering what The Document Foundation has been pushing for lately, a locked-down proprietary-only shortcut feels like an odd choice.

Installing LibreOffice and removing ONLYOFFICE didn't fix it either. The shortcut still generates files in those same formats; it just opens them in LibreOffice instead.

The file manager looks different too, even though it's the same GNOME Files under the hood. When you launch it, the app opens a "Computer" page that lays out your folders, disks, and removable media in one clean view.

The default app selection is generous. You get ONLYOFFICE, Gear Lever, Warp, Mousam, and a good spread of the usual GNOME-focused apps alongside them.

stillOS also ships its own lineup of "still" branded tools. stillTerminal is the default terminal app, built around a customized ZSH setup with container and SSH support.

stillCenter is the curated app store, pulling from Flatpak and other sources with its own stillRating system grading how well an app runs with stillOS.

stillControl handles GNOME layout customization, letting you switch between 12+ preset layouts and manage extensions without digging through GNOME Tweaks. Then there's stillOS Forums, which isn't a native app at all. It's the community forum wrapped as a SWAI web app, so it behaves like part of the desktop instead of a browser tab.

Running other software worked without any surprises. I selected Zen Browser during onboarding specifically to see how it handled a non-default browser, and it ran cleanly, without any visual bugs, lags, or crashes to speak of.

The pill-shaped workspace switcher held up well through all of this too, letting me flip between workspaces and open apps without any lag or fumbling. Though this was a given seeing that it is a well-tested component from GNOME.

And before you assume that this is some obscure Linux distro that will fade into irrelevance very soon, know that development is still quite active. The recent stillOS 10.2 release introduced most of what's covered above.

Cameron says that:

stillOS has now reached the point where it fits that initial vision. There is obviously more work to do, and there always will be, but the foundation is here. Now I want to focus on building the community around it.

The goal is to keep stillOS innovative while also building a Linux system that literally anyone can use. Those ideas do not have to conflict. Linux can try new things without becoming confusing, and it can be approachable without becoming boring.

Download it now

You can grab the latest stillOS ISO from SourceForge, which lists all the images released by the project so far.

Once you're up and running, know that updates aren't handled through AlmaLinux's usual dnf-based flow.

Instead, you can rely on the distro to handle it for you, as it checks for updates automatically 15 minutes after boot, then every 24 hours after that for as long as your machine stays on.

If you can't wait, then you can run the following command in the terminal to force an update check:

sudo bootc upgrade

Xen is a hypervisor that runs on various kinds of hardware, where a crash can have massive consequences. It is deployed in fields like automotive, industrial automation, robotics, and avionics. These are all safety-critical fields where certification matters.

So far, it has been on a good trajectory with increasing adoption and enhancements that keep it relevant for enterprise use. What it hasn't had is a structured, collaborative way to help the companies building on it actually get certified for that kind of work.

It has launched the Xen Safety Committee, a group tasked with building the safety documentation and test evidence that member companies can lean on instead of starting from scratch.

The committee's job covers software safety requirements, architecture specifications, MISRA-related engineering, DFMEA analysis, testing frameworks, tooling, code coverage, and process documentation.

Founding contributors AMD, EPAM, and Renesas are credited with doing most of the legwork for the above, forming the initiative's early foundation.

The point is to stop every company from solving the same problem on their own. Sharing this foundation should cut costs, avoid duplicated effort, and get products through certification faster for anyone building on Xen.

Nothing changes about how Xen itself gets built. Source code, licensing, and the existing open contribution and review process stay exactly as they are. The Safety Committee runs as a parallel governance layer, not a replacement for how maintainers already work.

Though joining the committee isn't all that open. Organizations will need to sign up for the right kind of membership. More on that below. 👇

A call for new members

Membership at Xen Project works like most Linux Foundation projects. It funds shared infrastructure, test environments, events, and coordination, while day-to-day technical decisions stay with maintainers through open contribution.

Paying to join doesn't buy influence over the codebase. 🤓

AMD, ARM, AWS, Boeing, EPAM, Ford, Honda, Renesas, Vates, and XenServer are all listed as current members. These represent chipmakers, cloud providers, automotive, aerospace, and a few other sectors.

A new membership tier, "Premier Plus" sits above the existing tiers, granting voting representatives on both the Xen Advisory Board and the Safety Committee, a say in setting priorities and roadmap, and access to the committee-managed safety artifacts.

This way, instead of producing every requirement and test in house, members can hand the committee's existing artifacts to their own safety assessors for audits and assessments.

Speaking on the topic, Kate Stewart, Vice President, Dependable Embedded Systems, The Linux Foundation, stated that:

Systems with safety considerations need a trusted hypervisor to provide the foundation for other open source components, such as Zephyr and Linux, to support mixed criticality embedded applications.

The Zephyr and Xen projects have been exchanging best practices over the last few years as a pathway to achieve safety certification of open source components. We welcome the Xen Safety Initiative to accelerate the journey of Xen achieving formal certification.

Suggested Read 📖: Proxmox Virtual Environment Officially Runs on ARM64

One thing that always comes up in the GNOME vs KDE debate is how easy it is to customize KDE. You can surely customize GNOME but not to the extent of KDE.

But that doesn't deter us GNOME users from trying new looks and visuals, does it? In that regard, I recently came across a super cool GUI tool that lets give an altogether different look to your GNOME desktop.

Take a look at the screenshot below. Looks like Archcraft or Omarchy, right?

That is the magic of Rewaita, a GUI app worth adding in your customization toolbox.

Rewaita: Add custom color tint to GNOME Adwaita

Rewaita is not a standalone theme or icon. It keeps GNOME’s familiar Adwaita design but replaces its colours and adds optional visual effects like changing the transparency to windows.

As you open Rewaita, it very helpfully gives you a user guide to help you through the process of changing the theme on GNOME.

Once you're through that process (that I'll describe later), it gives you a good range of choices that include 14 light themes and 24 dark ones (all hail dark mode), followed by 6 Window Control themes. Many of these themes are well-known standard themes around Linux, like Breeze, Nord and Catppuccin, etc.

Which is great because they're all tried and tested to suit the popular tastes of Linux users. After all, most Linux ricing screenshots you see on various communities use those color themes.

The Window Control themes are more of the same that include Breeze, Mint and MacOS-styled button themes.

All you need to do is choose the theme, and the change is immediate in all newly opened windows.

Under the hood, it generates CSS overrides and related assets inside directories such as

~/.config/gtk-3.0/
~/.config/gtk-4.0/
~/.local/share/themes/

Note that it will only affect GTK applications. Qt (from KDE) and Electron apps may not adhere to the changed color theme.

Using Rewaita

To activate Rewaita, you have to ensure a few things first.

Make sure you have the gnome-shell-extensions package installed, which allows you to extend the capabilities of GNOME.

And then the User Themes extension (I would also recommend installing the Extension Manager, that makes it easier and more fun to explore various extensions that GNOME has to offer), as well as the GNOME Tweaks application (where you actually choose the theme).

When you're done with all that, finally install Rewaita, which is available on Flathub. If you prefer terminal, here is the command:

flatpak install io.github.swordpuffin.rewaita

Other ways of installation on distributions such as Arch, OpenSUSE, and NixOS are listed here.

After that, go into Rewaita and choose a theme. Once you do that, reboot your system, go into GNOME Tweaks and choose Rewaita as the Shell theme.

Flatpak Override

Since Flatpak applications are treated separately sandboxed, these settings are not applied to them by default. You can, however, make them apply by overriding the settings:

sudo flatpak override --filesystem=xdg-config/gtk-3.0:rw
sudo flatpak override --filesystem=xdg-config/gtk-4.0:rw

The first command takes care of GTK-3.0 applications, the second does that for GTK-4.0.

More customization options

If you desire a more manual and methodical approach, you can even choose the exact colors that you want there to be on the interface on the Custom tab.

Accent colors, interface colors, button colors, successful, dangerous or warning action colors, etc. can all be manually chosen, alongwith manual color naming and extra CSS code. You can even save these settings as themes on your system for future use.

Fine Tuning

The Fine Tune section is another wonderful aspect where you can put in use some extra settings apart from the colors on your system. These include things like transparency, window borders, sharpening corners, etc. There's an option to toggle Rewaita running in the background in case you swap between light and dark modes throughout the day.

Firefox

You can choose to apply the theme to your Firefox window as well (great edge for the full open source application pack) by enabling 'Generate Firefox CSS Theme' in the Fine Tune tab. After that, Firefox only needs to be fully quit and restarted for the settings to apply.

Wrapping Up

Rewaita is really, really, really convenient. It makes it extremely easy not to only apply pre-made themes, but also create custom themes of your own with colors and CSS code.

More options are just a toggle switch away, and the user guide that is present right as you open the application is extremely helpful, telling you all that you need.

Thoughtfully constructed and beautifully applied, Rewaita is a must use for anyone wanting to make GNOME look aesthetically pleasing.