@itsericwoodward@itsericwoodward.com

and/or the people in charge realize that “enterprise-class” and “vibe-coded” are mutually exclusive software concepts

I’m really uncertain about this, to be honest. Short-term, you can vibe-code a ton of stuff and get away with it. I think that quality/correctness only matters for a small fraction of software out there. But long-term? I guess that then depends on whether you’re able to still pay the AI bills. Maintenance/refactoring is going to be veeeeeeeeeeery expensive, and virtually no human will be able to do those tasks. Or, you know, just rip off your customers and don’t sell any long-lasting support contracts in the first place.

I dunno. 🤷‍♀️

I hope things get better for you, @movq@www.uninformativ.de (and for anyone else dealing with the same issues).

Thanks! 🙏

@lyse@lyse.isobeef.org

The IT industry is absolutely poisoned and destroyed.

Yes, and here’s an example:

• https://hachyderm.io/@thomaswilburn/117439476156199985
  
• https://chitter.xyz/@gabu/117441628277548428
  

I’m seeing exactly this effect in our stuff at work.

And human PR reviewers (in our company) are already completely burned out, because the AI dudes just dump all their shit on us. AI is an amplifier: If you already didn’t care about good code before AI, then you most certainly won’t start caring with AI, so the result is just a massive dumpster fire.

I suppose you can use AI to produce good code, but this takes so much time, effort, and money, you might as well code it yourself.

Yeah, things aren’t quite that bad where I work yet (thanks to no one wanting to pay the outrageous bills associated with the “frontier models”), but some people keep pushing for it. Personally, I’m still hoping that either the bubble bursts real soon and/or the people in charge realize that “enterprise-class” and “vibe-coded” are mutually exclusive software concepts, but who knows.

In the meantime, at my work, we’re still doing mandatory human PR reviews (with optional copilot ones), and there’s a big push for “respectful” AI use (as in being respectful of others, like always reading any text / code the AI generated for you before asking someone else to). We’ll see how it goes.

I hope things get better for you, @movq@www.uninformativ.de (and for anyone else dealing with the same issues).

Epic Games’ New Vibe-Coded Launcher Is 46% Faster Than Steam
Epic has launched a preview of a new AI-generated Games Launcher that it says cold-starts 46% faster than Steam. Neowin reports: Epic commissioned a third-party company to benchmark this new launcher in its current Preview build versus the Steam client on Windows machines. The results reported that Epic’s client is 46% faster than Valve’s client during co … ⌘ Read more

Apple Contributes SSAF Driver To LLVM Clang For Source Code Transformations
Apple compiler engineers have contributed to upstream LLVM/Clang today a Scalable Static Analysis Framework (SSAF) driver for whole-program based analysis and source-to-source code transformations… ⌘ Read more

Google Play Now Lets Anyone Build Private, Vibe-Coded Apps That Run On Android
Google Play is adding support for private personal Android apps, allowing users to build an app with tools like Google AI Studio and run it on their own devices without publishing it publicly to the Play Store. “These personal apps won’t show up in Google Play’s search or listing pages, and won’t be shared with anyone e … ⌘ Read more

Anthropic Reinvents Source Control As Claude Floods the World With Code
Anthropic says Claude-generated code has scaled so quickly that its existing Git-based infrastructure could not keep up. In response, the company built a custom centralized source-control system called Scion, along with AI-assisted review and search tools. The Register reports: “In May 2026, Anthropic averaged around 2,500 cod … ⌘ Read more

AI usage at work has risen to absurd and moronic levels.

Entire dev teams getting dismantled and fired.

The people that work in accounting are expected to write their own accounting software (or rather, let AI do it). Same for marketing. And everybody else.

Former “tech nerds”, whom I used to respect, tell me that they haven’t read any code in months. Let alone written some. They tell me that “AI is really good now”, it’s the future, yadda yadda. At the same time, when I read their code, I want to scream and cry (and most days I do).

Virtually everything is vibe-coded now. Everywhere you look you get to see code that no human will ever be able to maintain. Huge system “designs”, expensive, slow.

At this point, I’m almost hoping for the bubble NOT to burst, because that might very well mean the end of this company (and, thus, my job).

There is no “exit strategy”. We completely depend on AI services now.

At the same time, I’m expected to implement ISO27001.

Everything is vague, chaotic, wonky, shitty, buggy.

People are actively working on making themselves obsolete, and they don’t seem to notice that they’re doing that, because they’re so engulfed with playing with their new AI toy.

I checked job offers, but it looks like every company is like this now. Maybe some are still a bit behind, but sooner or later, they’ll all be like that.

And the worst part is: None of this matters in the slightest, as long as there’s some idiot out there who buys our stuff. Welcome to capitalism.

People keep telling me that you can “use AI responsibly” and “just to some degree, just as a tool”, but they don’t seem to realize that this does not matter in capitalism. Companies don’t need “good products”. They just need to sell something. If you still believe that “doing a good job” matters and that “they still need competent people”, haha, my sweet summer child, nothing you do matters. You just happened to be needed in the past, but those days are gone.

The future is absolutely bleak – assuming there is a future.

Google Launches Flagship AI Model Gemini 4 Argon
An anonymous reader quotes a report from Reuters: Alphabet’s Google on Wednesday released its flagship AI model Gemini 4 Argon to developers and businesses worldwide, seeking to regain ground against rivals OpenAI and Anthropic after months of delays and a leadership shakeup at its DeepMind lab. The model excels at “deep work” on complex, multi-step tasks in coding, finance, la … ⌘ Read more

IREE 3.12 Introduces Rolls Out Rewritten Vulkan Driver, Experimental AMD HSA HAL
Rolling out last week was a new version of IREE as the Intermediate Representation Execution Environment for MLIR-based end-to-end compiler and run-time for lowering machine learning models into a unified IR and working across different hardware/driver stacks. Making IREE 3.12 notable is a rewrite to its Vulkan code… ⌘ Read more

AI Coding Agents Generate More Code, But Not More Software
A study of more than 700 software firms found that AI coding agents significantly increase the amount of code developers produce, but not the amount of software companies actually ship. Ars Technica reports: A recent study (PDF) of actual coding practices across hundreds of firms finds that human code review forms a significant “bottleneck” for the overall ef … ⌘ Read more

Linus Torvalds “Did What I Preach”, Used AI To Make Code Decipherable & Uncovered More Bugs
After using Claude Code to draft code to help make a fundamental improvement to the Linux kernel, Linux creator Linus Torvalds continued using AI this weekend and talking up its benefits. Beyond generating initial patches and a test program for that noted I/O enhancement, he separately talked up using AI/LLMs for making patches more decipherable to understand in the process the AI noted more long-standing bugs within the … ⌘ Read more

Intel Posts Updated Linux Driver Code For HDMI 2.1 FRL On Meteor Lake & Newer
Following AMD completing their HDMI 2.1 implementation for the open-source AMDGPU Linux driver after beginning with HDMI FRL patches earlier in the year, Intel’s open-source driver developers are following in their footsteps… ⌘ Read more

Linus Torvalds Working On A Hefty Linux Kernel Improvement With Claude Code
Linux creator Linus Torvalds has been known to use AI to help with kernel debugging and outside of the kernel tree has also done some vibe coding with Google’s Antigravity. He has previously said he views AI as a useful tool. It turns out he’s even now using Claude Code for working on new Linux kernel features… ⌘ Read more

Apollo Software Pioneer Margaret Hamilton Dies at 90
“Margaret Hamilton, the groundbreaking researcher who helped create the field of software engineering and whose code helped land NASA’s Apollo astronauts on the moon to win the 20th century space race, has died,” writes longtime Slashdot reader Veeru. “She was 90.” MIT News reports: A computing pioneer who authored over 130 publications, Hamilton helped to establish … ⌘ Read more

Secretive Rosaic Labs Hires Legendary Linux x86 Developer
The curiosity into secretive semiconductor startup Rosaic Labs builds. Earlier this year it was noted that Intel was providing Rosaic Labs with access to its Atom technology with RTL code. There have also been rather mysterious x86 additions not from Intel or AMD (or Zhaoxin / Hygon) that given the timing may pertain to Rosaic Labs. News today is that Rosaic recently hired a long-time, Linux x86 developer veteran… ⌘ Read more

Google’s Sashiko AI Has Completed 191k Patch Reviews, Cited On Nearly 500 Kernel CVEs
Sashiko that was built by Google engineers for agentic review of the Linux kernel code changes and powered by Google’s AI models continues proving very valuable to kernel developers. In less than one year since they launched Sashiko for agentic AI code review, their results are quite fascinating… ⌘ Read more

AI Makes “Open-Source, Clean Room” Implementations Of Adobe Photoshop & Premier In Rust
As an interesting demonstrator for the power of AI and potentially pressing legal challenges, Claude Code and AI agents have created “open-source, clean-room reimplementation” versions of Adobe Photoshop, Adobe Premier, and Adobe Lightroom. These new creations from the ArtCraft project are implemented in the Rust programming language… ⌘ Read more

Qualcomm Posts QPaCE Linux Driver Code For Faster ZRAM On Newest SoCs
Qualcomm’s Snapdragon 8 Elite Gen 6 “Hawi” SoC features the Qualcomm Page Compression Engine “QPaCE” hardware accelerator. Linux patches posted recently provide a QPaCE driver for accelerating page compression and decompression for the ZRAM block device… ⌘ Read more

IBM and Red Hat Find More Than 400 New Vulnerabilities In Popular Java Code
IBM and Red Hat say their AI-powered Lightwell initiative has identified and helped remediate more than 400 previously unknown vulnerabilities across widely used Java libraries. Phoronix reports: They announced this feat today as part of their promoting Lightwell Clearinghouse to GA, which is an enterprise service for their cu … ⌘ Read more

IBM & Red Hat Find More Than 400 New Vulnerabilities In Popular Java Code
IBM and Red Hat announced today they have identified and remediated more than 400 previously unknown vulnerabilities across popular Java libraries… ⌘ Read more

YSERVER 1.7 Rust X11 Server Narrows Memory Usage Difference With X.Org Server
YSERVER continues seeing a lot of activity for this vibe coded X11 server written in the Rust programming language. YSERVER has been seeing many apps and desktops working fine on this creation crafted in large part by Claude Code and continues chipping away at more X11 features. YSERVER 1.7 is out today with the latest fixes and improvements… ⌘ Read more

Everything about time management is a pain for developers: time zones (and the locales that ignore them or go half-way), leap seconds, all of it is messy and weird and human. And DST is the worst aspect of it by far.

My first professional coding job as part of a team was in a Java shop, and there was a running joke along the lines of “Only Java would have a class called SimpleDateFormat which is anything but simple.”

Linux’s ZRAM Reworked For Greater Memory Savings, Better Performance
With today’s very high memory prices, the Linux ZRAM code for compressed block devices in RAM has seen a rework to provide greater memory savings… ⌘ Read more

System76 COSMIC projects will no longer accept LLM-generated content in code submissions
The AI machine continues to cause problems for some developers - with System76 now banning LLM-generated content in code Pull Requests for COSMIC.

Image

Read the full article on [GamingOnLinux](https://www.gamingonlinux.com/2026/10/system76-cosmic-projects-will-no-longer-a … ⌘ Read more

OpenCourant: Rocky Linux Developers Create Community Fork Of OpenRadioss
This week was the surprising and unfortunate decision of Siemens shutting down the OpenRadioss project as the four year old open-source project started by Altair Engineering with their prominent Radioss finite element solver. Siemens didn’t just end the project but they shutdown the GitHub repository that hosted the open-source code and removed other resources that had built around it. Fortunately, there’s a new community fork of OpenRadioss … ⌘ Read more

Mesa Drops Its Poorly Maintained & Rarely Used Direct3D 10 User-Mode Driver
Back in 2021 the Direct3D 10 User-Mode Driver “d3d10umd” was introduced as a new Gallium3D state tracker to work as a Windows WDDM UMD driver similar to Windows WARP while leveraging LLVMpipe for better performance. Now five years later the poorly maintained code is removed from the Mesa codebase… ⌘ Read more

Some Networking Fixes Being Diverted To Linux 7.4, AI/LLM Activity Still Increasing
With the vast array of wired and wireless networking drivers within the Linux kernel and range of protocols and other networking features, it’s a big expanse of code for AI/LLM agents to analyze and critique. The Linux networking developers have acknowledged being “complete overwhelmed” by the AI/LLM-driven patch activity and bug reports as well as battling against AI slop patches. Even though Linux 7.3 stable won’t be out f … ⌘ Read more

Clearing Out Linux’s Old 32-bit ARM Platform Code Is Itself A Challenge
Longtime ARM Linux maintainer Arnd Bergmann over the past year has been working to clear out a number of old ARM platforms from the mainline kernel tree. With Linux 7.3 many older 32-bit ARM platforms were deprecated and in turn hundreds of kernel drivers orphaned. Now it’s on to removing that actual code, which given the amount of entangled code, is itself a challenge… ⌘ Read more

NVIDIA Olympus SMT Optimizations For Vera CPU Showing 3~6% Performance Gains
Back in August I wrote about NVIDIA working on Linux scheduler improvements for helping the SMT performance on their new Vera CPU with Olympus cores. That work remains ongoing but is looking fairly beneficial for performance especially for knocking it out in less than one hundred lines of code… ⌘ Read more

@movq@www.uninformativ.de Oh, it’s been a hot minute since I last came across Pascal. :-) When I first quickly skimmed through the code, my eyes fell on the reproducer in the comment and I immediately thought: This reminds me of Delphi! And sure enough, I then saw the .PAS file extension in the URL. That Kiwi applied some strange code formatting, I have to say, though.

Haha, that’s a sweet trick with stdin. Crazy, who comes up with things like that? It would have never crossed my mind to do stunts like that. :-D

Today, I also learned that -R is an alias for -r in rm. I think I never saw -R used here before. Of course for chmod and chown, but not rm.

NOCL Making Progress As Mesa Gallium/OpenCL Atop NVIDIA CUDA
In development the past year has been the Mesa Gallium3D API implemented atop NVIDIA’s CUDA driver API. Initially named “CLUDA” and now going by “NOCL”, this Mesa code paired with Rusticl is working well for OpenCL atop the NVIDIA proprietary driver as a new compute option… ⌘ Read more

Mesa 26.3 Enables Intel’s New Jay Shader Compiler By Default, Compile Times 55% Faster Or 66% Faster Than Windows
Over the past year Jay has been in development as a new open-source shader compiler for Intel GPUs. Alyssa Rosenzweig has been leading its development since joining Intel in 2025. For Mesa 26.3 the Jay shader compiler is being enabled by default for modern Intel graphics hardware and is around ~10% faster than the existing BRW compiler code… ⌘ Read more

“Shiva” Being Developed For Mesa Vulkan Drivers, Akin To Gallium3D
Shiva is entering development for Mesa Vulkan drivers as a new framework akin to Gallium3D but for Vulkan drivers rather than OpenGL. It’s going to be some time before the work will come to fruition but should hopefully lead to more code sharing and a better Mesa Vulkan driver development experience in the future… ⌘ Read more

Legacy Code is a bizarre looking mechanical creature action-physics sim
Legacy Code is one of the more quirky games to enter the GamingOnLinux inbox recently, where you build up a mechanical creature in a action-physics sim.

Image

Read the full article on [GamingOnLinux](https://www.gamingonlinux.com/2026/09/legacy-code-is-a-bizarre-looking-mechanical-creature-action-phy … ⌘ Read more

FEX code cache enabled for Proton Experimental (ARM) to improve frame timings
Valve have upgraded Proton Experimental (ARM) with a feature that should hopefully improve some frame timings for devices like the Steam Frame.

Image

Read the full article on [GamingOnLinux](https://www.gamingonlinux.com/2026/09/fex-code-cache-enabled-for-proton-experimental-arm-to-improve-frame- … ⌘ Read more

Linux Kernel’s LZ4 Compression Code Being Resynced For Better Performance & Cleanliness
In addition to the Linux kernel’s Zstd compression code being improved, the LZ4 compression code within the kernel tree is also seeing a separate set of enhancements… ⌘ Read more

Hey, folks. Sorry to disappear for the last couple of weeks.

Last weekend, I took the family to Atlanta, GA for a visit to the Georgia Aquarium and the Center for Puppetry Arts, both of which were fantastic (pics forthcoming). Unfortunately, amongst the souvenirs we brought home was the worst sickness my wife and I have had in years, which basically laid us both out until Thursday, and we’ve been recovering since.

On top of it, just before we left, I accidentally broke TwtStrm (my WIP TwtTxt reader), so I didn’t even have a functional way to keep up with the only “social media platform” I actually give a hoot about (nor a functional brain to try and fix it).

As of this morning, though, I’m happy to report that we’re basically all better, and TwtStrm is back to working. I’m actually hoping to get some code written this weekend (and maybe even try to get Parley running), but we’ll see how things go. 🙂

ReactOS Lands Significant DirectX Stack Update From Wine For Enabling More Games
The open-source ReactOS project striving to provide binary compatibility with Microsoft Windows applications/games as well as drivers took a huge step forward today with landing a significant update to its DirectX stack from code leveraged from the Wine project… ⌘ Read more

“slab_tiny” Boot Option Proposed For Tiny Linux Systems With Very Limited RAM
In 2022 the SLOB allocator was deprecated and removed with Linux 6.4 a year later. SLOB was popular with embedded systems with minimal amounts of RAM, so as a result the CONFIG_SLUB_TINY Kconfig option was then added for configuring the slab allocator for a minimal memory footprint for systems with 16MB of RAM or less. The CONFIG_SLUB_TINY is now on the chopping block with a proposed rework to the allocator code for making the “ … ⌘ Read more

Linux 7.3 Reverting Logitech Bolt Receiver HID++ After Too Many Bugs Encountered
Merged at the start of the Linux 7.3 kernel cycle was adding Logitech Bolt USB receiver support for Logitech HID++ devices to the hid-logitech-dj driver code. That has resulted in too many bug reports of early Linux 7.3 users encountering varying issues with their Logitech mice that the functionality is now being reverted until it can be better developed… ⌘ Read more

Are Students Suddenly Losing Interest in Computer Science as AI Coding Takes Off?
On academic mentoring platform Nova Learning, Computer Science and AI, “once the dominant choice among students… is losing ground fast, while Engineering has nearly doubled its share.” It’s a small survey, but “The steepest proportional decline is in Software & Data Science, the most traditional ‘learn to code’ … ⌘ Read more

systemd 262 Released With AI/LLM Canary For Unreviewed Code, LUO Integration
Systemd 262 made it out today as the newest feature update to this Linux init system and service manager. As with most systemd releases, it’s heavy on new features and functionality… ⌘ Read more

Vibe Coded YSERVER Adds XDMCP Support For Multi-User Shared Desktops
The open-source YSERVER project as a modern X11 sever written in the Rust programming language with the help of Claude Code is out today with a new feature release. Most notable for this AI Rust X11 server is X Display Manager Control Protocol (XDMCP) support for multi-user shared desktops… ⌘ Read more

Gravity Linux Aims To Provide A Nice Experience For Apple M4 & Newer
There’s a new Linux distribution focused on supporting Linux on Apple Silicon Macs. Gravity Linux is this new distribution that is a spin of Fedora Remix and is catering to the still-experimentally-supported Apple M4 Macs and newer. In addition to leveraging Asahi Linux code and reverse engineering, Gravity Linux also leverages OpenAI Codex and Claude Code… ⌘ Read more

Linux 7.3-rc4 Released: More Fixes Caught By LLMs, But Nothing Too Scary
Linus Torvalds announced the release today of Linux 7.3-rc4. As usual for recent months, the fixes continue to be quite heavy and scattered all over the place driven in large part by AI/LLMs spotting various code defects. Overall though Linus Torvalds isn’t too worried about Linux 7.3-rc4… ⌘ Read more

AI-Assisted Starling Desktop Introduces 3D View With “Walkable 3D Waterfront City”
From a few months ago you may recall Starling as the new Linux desktop written with its own Wayland compositor, using the Swift programming language, and written in large part via AI with Claude Code. Out today is Starling 0.5 with this desktop now gaining an optional 3D desktop mode… ⌘ Read more

Apple Video Decode Driver Posted For Upstream Linux Kernel Review
The latest Apple Silicon enablement code working its way from the Asahi Linux project toward the upstream Linux kernel is the AVD driver used for accelerated video playback across Apple M1 / M2 / M3 SoCs… ⌘ Read more

Rust is Now a ‘Tier One’ Language at Microsoft
The Register reports:

When Microsoft’s developers and engineers sit down to code, they can now choose to work in Rust, which Redmond has added to its list of canonical languages. “Rust now is a Tier One language at Microsoft, and that just means that it sits among C++, C# and TypeScript as the best supported languages for internal development in the company,” explained Victor … ⌘ Read more

Linux Kernel Build Times Ready To Be Significantly Reduced With Latest Patches
The past few weeks we have eagerly been following a Linux kernel patch series after AI found some major bottlenecks in the Linux kernel build code. With AI helping to address these bottlenecks, Linux kernel build times are substantially reduced. These optimizations might be ready in time for Linux 7.4 for reducing incremental kernel build times as much as around 70%. This week the latest iteration of these patches were posted by Arm engi … ⌘ Read more