Linux

These platforms help teams monitor how technical content appears in AI-generated answers, from brand mentions and citations to accuracy and referral traffic.

Search is splitting into two experiences: one built around ranked pages and another around generated answers. As AI platforms summarize Linux tutorials, open-source project documentation, API references and technical guidance, teams need ways to measure whether their content is mentioned, cited and accurately represented.

An AEO platform provides that visibility. They help writers, maintainers, developers, and content teams examine how pages appear in generated answers, which sources receive citations, and whether AI platforms send traffic to the intended documentation.

Top AEO Tools for Technical Content

AEO and AI visibility platforms vary greatly. Some are mostly about tracking mentions, while others combine AI visibility with citations, referral traffic, competitor data or content recommendations. Below is a selection of tools, both popular and niche, that can be used by technical and content teams responsible for developer portals, Linux resources and open-source projects.

Best for: Measuring AI visibility, citations and resulting website traffic in one platform.

Similarweb AI Search Intelligence gives teams a broad view of how brands and websites appear across AI-generated search experiences. It can be used to monitor brand mentions, citation frequency, prompt-level visibility and share of voice while comparing results with competitors. Similarweb also connects those measurements with its wider web and traffic intelligence, helping teams examine whether visibility and citations are producing visits to particular pages.

For technical content teams, this combination can make it easier to trace the path from an AI answer to its cited source and then to referral traffic. Teams can identify prompts where documentation is absent, determine which domains are being cited instead, and investigate whether AI-referred visitors reach the intended technical

pages. For example, an open-source project could examine whether prompts about installation on Ubuntu, package dependencies or command-line configuration lead users to its current documentation rather than an outdated forum post.

Standout feature: Similarweb combines AI visibility, citation analysis and traffic intelligence, reducing the need to examine those signals in separate platforms.

Best for: Tracking technical content across a wide selection of AI engines.

AI security means hardening everything around the model, not just the model itself.

AI workloads have moved out of the research sandbox and into production, which means they now carry the same operational weight as any other critical service. If you want to understand what is AI security in practical terms, start here. It's not a 

separate discipline bolted onto your existing stack. It's an extension of it. The systems around a model matter as much as the model itself, and for teams running Linux and open-source infrastructure, that means applying familiar discipline to some new risks.

AI Security Is Becoming Part of the Linux Infrastructure

Treat AI workloads like production infrastructure, not side projects. A model doesn't run in isolation. It depends on an operating system, a container runtime, storage, APIs, identity systems, and a pile of dependencies. If any of those layers are weak, prompt filtering or output moderation won’t be enough.

Security teams that only look at the model may miss much of the actual attack surface. The work starts with the same questions you'd ask about any service. What's exposed? Who can reach it? What happens if it's compromised? 

The AI Attack Surface Extends Beyond the Model 

Model-serving APIs, embeddings, retrieval systems, vector databases, plugins, and CI/CD pipelines all now sit inside the perimeter. A vector database that holds proprietary documents is a data store like any other and needs the same access controls. An inference endpoint that accepts arbitrary text input is a public-facing service, and it needs rate limits and authentication like any other. Plugins and tool integrations expand what a model can touch, which also expands what an attacker can touch. It’s the same threat modeling you’d apply to a web application. 

Why security buyers are rethinking what matters most.

Security teams typically don’t struggle to find vulnerabilities as much as they have in the past. The harder part usually begins after the report arrives, once dozens of findings land in front of engineering teams already juggling patch schedules, production deadlines, and internal disagreements about urgency. Platforms like XBOW, OffSec, and Cobalt have entered that environment as organizations started rethinking what they actually need from pentesting vendors beyond annual compliance exercises.

A vulnerability may look severe inside a dashboard, while no one internally agrees whether it creates meaningful exposure or simply adds another item to an already crowded queue. Infrastructure also changes too quickly for static testing cycles to answer every operational question.

APIs update mid-quarter, contractors receive temporary access that lingers longer than expected, and cloud permissions change quietly during routine development work. Buyers evaluating vendors now spend more time asking whether testing accurately reflects the systems employees use every day.

Pentesting Vendors Now Face Different Expectations

Long reports stopped carrying the same weight years ago. Security teams already know modern environments contain weaknesses. What many organizations want now is clearer evidence showing which findings deserve immediate attention and which ones can wait without creating major operational exposure. That distinction became harder to ignore as remediation timelines stretched across larger environments.

Linux kernel developers are considering a new “steal governor” designed to improve performance when multiple virtual machines compete for limited physical CPU resources. The proposal uses the amount of CPU steal time observed inside a guest to dynamically reduce or expand the number of virtual CPUs on which that VM prefers to schedule work.

The feature is primarily aimed at heavily virtualized servers where administrators deliberately assign more virtual CPUs than the host can physically execute at once. Under heavy load, that overcommitment can lead to frequent vCPU preemption, lock-holder delays, cache disruption, and ultimately lower overall throughput.

The latest v11 patch series was posted on August 25, 2026, and its developer has proposed consideration during the Linux 7.3 development cycle, potentially targeting Linux 7.4 for inclusion. This means the feature is still under review and is not part of a stable Linux kernel yet.

What Is CPU Steal Time?

CPU steal time is a concept specific to virtualization.

Imagine a virtual machine has eight vCPUs. From inside that VM, the operating system behaves as though those eight CPUs are available. But those virtual CPUs ultimately need to run on the host's physical processors.

If several VMs are competing for the same physical CPU resources, the hypervisor may temporarily prevent one VM's vCPU from running so another VM can use the processor.

The time during which the guest wanted to execute but couldn't because the hypervisor was using the underlying CPU elsewhere is known as steal time.

High steal time is therefore a useful indication that the physical host is experiencing CPU contention.

The “Noisy Neighbor” Problem

The steal governor is designed primarily to address what virtualization engineers commonly call the noisy neighbor problem.

Consider a server hosting several VMs:

• VM A has 32 vCPUs.
• VM B has 32 vCPUs.
• VM C has 32 vCPUs.
• The physical server has only 64 CPU threads available to those workloads.

That configuration can work perfectly well when the VMs aren't simultaneously busy.

If all three suddenly become heavily loaded, however, they may collectively request more CPU time than the physical machine can provide.

The hypervisor then has to constantly switch between vCPUs.

Those interruptions can become particularly expensive if a vCPU is preempted while holding a lock or executing another latency-sensitive section of code. Other threads may then wait for a vCPU that isn't currently being allowed to run.

The result can be counterintuitive: giving the VMs more virtual CPUs can sometimes make the combined workloads slower.

The Thunderbird team has officially released Thunderbird 154, delivering several useful new features alongside a substantial collection of fixes for email, calendars, address books, authentication, and stability. Released on August 18, 2026, Thunderbird 154 is now the latest monthly release of the popular open-source email client.

For Linux users, one of the most interesting additions is a new optional system tray mode, while Microsoft 365 users gain Microsoft Graph support. The release also improves global search, attachment handling, IMAP reliability, Exchange authentication, and CalDAV synchronization.

New Optional System Tray Mode

One of Thunderbird 154's most welcome additions is an optional system tray mode.

When enabled, closing Thunderbird's last window no longer needs to completely terminate the application. Instead, Thunderbird can remain running in the background through the system tray.

This can be particularly useful for users who want Thunderbird available throughout the day without keeping its main window open.

For Linux desktop users, the feature could make Thunderbird feel more like a traditional background email client, especially on desktops where tray-based applications remain part of the normal workflow.

Microsoft Graph Enabled for Microsoft 365

Thunderbird 154 also enables Microsoft Graph for Microsoft 365.

Microsoft Graph is Microsoft's API platform for accessing Microsoft 365 services and data. Its integration is particularly important as Thunderbird continues improving its support for Exchange and Microsoft-hosted email environments.

Thunderbird has been steadily expanding its Microsoft ecosystem compatibility, making the open-source client increasingly practical for users who need to access workplace Microsoft 365 accounts from Linux.

Global Search Gets a Useful New Option

Thunderbird's global search functionality also receives an improvement.

Users can now configure global search results to open in list view by default, providing another option for people who prefer a more traditional message-list workflow when reviewing search results.

It's a relatively small addition, but one that can make repeated searches more convenient for users managing large mailboxes.

Better Attachment Management

Thunderbird 154 introduces several improvements for handling email attachments.

A new "Copy To" folder context menu has been added for message/rfc822 attachments, and these attached messages can now be dragged directly into Thunderbird's folder tree.

These changes should make it easier to organize attached email messages without having to use additional steps or workarounds.

Linux creator Linus Torvalds has officially released Linux Kernel 7.2, opening another chapter in the development of the world's most widely deployed open-source kernel. The final release arrived on August 16, 2026, following seven release candidates and roughly two months of development. Kernel.org now lists Linux 7.2 as the latest mainline release.

Linux 7.2 is a substantial update with improvements spanning CPU scheduling, storage performance, AMD and Intel hardware support, graphics, networking, virtualization, RISC-V, and Apple Silicon. Among its biggest additions are Cache-Aware Scheduling, USB4STREAM, AMD ISP4 support, and initial AMDGPU HDMI 2.1 FRL functionality.

Cache-Aware Scheduling Lands in Linux

One of the most interesting performance additions in Linux 7.2 is Cache-Aware Scheduling (CAS).

Modern processors frequently contain multiple last-level caches shared between groups of CPU cores. Traditional scheduling decisions don't always account optimally for those cache relationships, potentially moving tasks between cores in ways that increase cache misses.

Cache-Aware Scheduling gives the Linux scheduler additional information about cache topology so it can make smarter decisions about where workloads should run.

Early testing during development produced particularly impressive results for some server workloads, including major gains in certain MySQL configurations. The improvement varies significantly by hardware and workload, so users shouldn't expect similarly dramatic gains everywhere.

For modern AMD and Intel systems with increasingly complex CPU topologies, however, CAS provides another tool for extracting better performance from existing hardware.

USB4STREAM Enables High-Speed Connections Between Computers

Linux 7.2 also introduces USB4STREAM, an interesting new capability for transferring data directly between computers over USB4 or Thunderbolt connections.

Rather than treating USB primarily as a traditional host-to-device interface, USB4STREAM can facilitate high-speed communication between systems.

Potential applications include:

• Large file transfers
• Development environments
• High-speed system-to-system networking
• Debugging
• Specialized workstation workflows

With USB4 and Thunderbolt increasingly common on laptops and desktops, this could eventually become a useful option for moving large amounts of data between Linux machines.

AMD ISP4 Finally Reaches the Mainline Kernel

AMD users receive another important addition with the arrival of the AMD ISP4 driver.

ISP stands for Image Signal Processor, hardware responsible for processing camera data before it reaches applications.

Linux kernel developers made a significant last-minute graphics change for Linux 7.2, reverting the DRM GPU scheduler back to its previous FIFO policy by default after the newer fair scheduling implementation caused serious performance regressions. The revert was submitted just before the final Linux 7.2 release after users reported severe slowdowns and desktop freezes under sustained GPU workloads.

The decision demonstrates the kernel community’s conservative approach to regressions: when a new feature causes problems this late in a development cycle, restoring known-good behavior can be safer than rushing an incomplete fix.

What Changed in Linux 7.2?

Earlier in the Linux 7.2 development cycle, the DRM subsystem switched its default GPU scheduling policy to a new fair scheduler.

The goal was to distribute GPU execution time more fairly between competing workloads. Instead of relying on the established FIFO scheduling behavior, the new implementation uses a fairness-oriented approach designed to prevent one workload from dominating GPU resources.

However, real-world testing exposed problems that had not appeared during earlier development.

The final DRM fixes for Linux 7.2 therefore:

• Revert the fair scheduler changes
• Restore FIFO as the default
• Restore selectable scheduling policies
• Mark the fair scheduling policy as experimental

DRM maintainer Dave Airlie acknowledged that developers may have moved too aggressively when switching to the new scheduler.

AMD Radeon Gaming Exposed the Regression

The problem became particularly noticeable during gaming on an AMD Radeon RX 9070 XT.

A regression report described severe performance degradation when the GPU remained close to 100% utilization. A game running through Proton could drop to roughly 10 FPS or freeze, while audio continued playing normally.

The problem could become severe enough that the entire KDE Plasma Wayland session stopped responding, sometimes requiring the compositor to be killed or the computer to be rebooted.

Interestingly, reducing GPU load—such as by opening an in-game menu or switching away from the game—could immediately stop the stuttering.

FIFO Worked, FAIR Did Not

Testing provided strong evidence that the new scheduling policy was responsible.

The reporter compared Linux 7.2 release candidates using FAIR against configurations using the previous FIFO policy. Under otherwise similar conditions:

FAIR: reproduced the performance problem.

FIFO: survived extended stress testing without the regression.

Linux 7.1.x also behaved normally.

Intel’s open-source Linux graphics stack has taken another step forward with hardware-accelerated AV1 encoding through Vulkan Video. New code merged for Mesa’s ANV Vulkan driver enables the VK_KHR_video_encode_av1 extension on Intel’s DG2/Alchemist graphics hardware, including Arc A-Series GPUs.

The development expands Intel’s Vulkan Video capabilities on Linux and gives applications another standardized way to access the dedicated video encoding hardware found in modern Intel GPUs.

AV1 Encoding Arrives in Intel ANV

The key change is support for the Vulkan extension VK_KHR_video_encode_av1 in Mesa’s open-source Intel ANV driver.

The extension was finalized by the Khronos Group in 2024 and provides a standardized Vulkan interface for hardware-accelerated AV1 encoding. It complements Vulkan Video’s existing AV1 decoding support and means Vulkan can provide both encoding and decoding interfaces for AV1, H.264, and H.265.

Intel had previously stated that its Arc graphics products would support Vulkan Video AV1 encoding through a future software update.

Initially Targeting Intel Arc Alchemist

The newly enabled Linux support specifically targets DG2, better known commercially as Intel’s Arc Alchemist GPU generation.

These GPUs already contain dedicated hardware capable of AV1 encoding, so the Mesa update does not add AV1 capability through software. Instead, it provides Vulkan applications with another way to access the GPU’s existing hardware video engine.

That distinction is important because hardware encoding can deliver much better performance and efficiency than encoding AV1 entirely on the CPU.

Why AV1 Matters

AV1 has become increasingly important for streaming, screen recording, video conferencing, and online video distribution.

The codec can provide high image quality at relatively low bitrates, making it attractive for applications where bandwidth and storage efficiency matter. It is also royalty-free, which has helped encourage adoption throughout the open-source ecosystem.

Hardware AV1 encoding can be particularly useful for:

• Game streaming

• Desktop recording

• Live broadcasting

• Video conferencing

• Video transcoding

• Content creation

For Linux users with supported Intel Arc hardware, Vulkan Video now has the potential to provide a common API for these workloads.

Building on H.264 and H.265 Support

The AV1 work follows recent improvements to Intel’s Vulkan Video encoding support for other codecs.

A newly proposed patch for the Linux kernel's AMD P-State CPU frequency scaling driver is showing impressive gaming performance improvements, potentially delivering a noticeable boost for Ryzen users without requiring new hardware. Early benchmarks indicate that the optimization can significantly improve frame rates in CPU-bound games by allowing processors to respond more quickly to changing workloads. (phoronix.com)

Although the patch has not yet been merged into the mainline Linux kernel, the initial results have generated considerable excitement among Linux gamers and kernel developers alike.

What Is AMD P-State?

AMD P-State is the modern CPU frequency scaling driver for AMD Ryzen processors on Linux. Rather than relying on the older ACPI CPUFreq driver, AMD P-State communicates directly with the processor to adjust clock speeds based on workload demands.

Its goals include:

• Faster frequency scaling
• Improved power efficiency
• Better responsiveness
• Higher performance during demanding workloads
• Lower power consumption when the system is idle

Most modern Linux distributions already support AMD P-State on compatible Ryzen processors. (kernel.org)

A Focus on Gaming Performance

The new patch specifically targets how quickly AMD P-State responds when a game suddenly demands additional CPU performance.

Many games rapidly alternate between light and heavy CPU workloads. If the processor takes too long to increase its clock speed, short performance dips can occur.

The proposed optimization reduces that delay, allowing the CPU to boost more aggressively when needed and helping maintain smoother gameplay. (phoronix.com)

Promising Benchmark Results

According to early testing, the patch delivers meaningful improvements across several Linux gaming workloads.

Reported benefits include:

• Higher average frame rates
• Better 1% low FPS performance
• Faster CPU frequency response
• Improved responsiveness during gameplay
• More consistent frame delivery

The biggest gains appear in CPU-limited games where processor performance has a greater impact than GPU performance. Systems that are already GPU-bound may see smaller improvements. (phoronix.com)

Designed for Modern Ryzen CPUs

The patch targets systems using the AMD P-State driver, which supports many recent Ryzen processors.

Compatible platforms generally include:

Linux kernel developers are moving forward with plans to remove the crypto_rng API layer, a long-standing component of the kernel's cryptographic subsystem. The proposed change is part of a broader effort to simplify the kernel's internal architecture by eliminating redundant code paths and encouraging developers to rely on the kernel's modern random number generation interfaces instead. (phoronix.com)

Although the change happens entirely behind the scenes, it reflects the Linux kernel community's ongoing commitment to reducing technical debt, improving maintainability, and modernizing core infrastructure.

What Is the crypto_rng Layer?

The crypto_rng framework is an API within the Linux kernel's Crypto API that provides random number generation services for kernel components.

Historically, it allowed different kernel subsystems and drivers to request random data through a generic cryptographic interface. Over time, however, the kernel's dedicated random number generator has matured considerably, making much of the crypto_rng abstraction unnecessary. (kernel.org)

Today, developers generally recommend using the kernel's built-in random number generation functions directly instead of routing requests through the older crypto layer.

Why Developers Want to Remove It

According to discussions on the Linux kernel mailing list, the crypto_rng layer has become largely redundant.

Modern kernel code already relies on well-established interfaces such as:

• get_random_bytes()
• get_random_u32()
• get_random_u64()

These functions are maintained as part of the kernel's primary random number generation subsystem and are widely used throughout Linux. Maintaining an additional abstraction layer increases code complexity without providing significant practical benefits. (phoronix.com)

Removing unnecessary infrastructure also makes the kernel easier to maintain and audit over the long term.

Simplifying the Crypto API

The Linux Crypto API has evolved significantly over the years as new algorithms, hardware accelerators, and security features have been introduced.

Kernel maintainers have increasingly focused on: