It was removed in Android 12
Markaos
The author acknowledges that, the blog post seems to be aimed at demystifying the concept of namespaces by showing that a "container runtime" that only does limited filesystem namespaces (using chroot) is enough to get some widely used containers running (of course without all the nice features and possibilities of the other types of namespaces)
saying that my legitimate copy of Windows 11 was at end of service
The screenshot says the version you use reached EoS and you need to update. There's absolutely nothing about invalid licenses in the screenshot.
Good job for getting upvotes on a "haha winblows bad" troll post, I guess.
As far as I was aware AMDGPU is used by default on most if not all distros
I really don't think that's the case, assuming you're talking about AMDVLK (amdgpu is the kernel module used by all three Vulkan drivers - RADV, AMDVLK and the Vulkan driver from AMDGPU-PRO). Ubuntu and Fedora definitely default to RADV, and Arch Wiki recommends RADV unless you need something from the other drivers.
I noticed a performance increase after forcing RADV on NixOS so not really sure.
NixOS seems to default to RADV according to their Wiki. If this was a few years ago then maybe you might be confusing it with the ACO shader compiler for RADV? That brought a significant performance increase and eventually became the default in RADV. I remember using custom Mesa (the project that develops open source graphics drivers, like RADV and radeonsi) builds to massively reduce stuttering in DirectX games.
I personally chose RADV after looking into this myself and the only drawback from my understanding is that they are proprietary drivers.
RADV is the open-source community developed Vulkan driver. It has the widest hardware support of the three Vulkan drivers and is generally the best for gaming.
AMD provides two more Vulkan drivers - AMDVLK is the open-source one available in AMDGPU, then there's the unnamed proprietary Vulkan driver in AMDGPU-PRO. The biggest advantage of the proprietary one is that it is certified - doesn't matter most of the time, but when it does, a missing certification is a deal breaker.
That depends a lot on how the license gets interpreted and how license violations are handled by the local law. The argument for why the end user cannot do anything about GPL violation is that the violated contract is between upstream and the "bad" developer - the upstream project gave the bad developer access to their source code under the condition that the license stays the same. You as the end user only get exposed to the bad developer's license, so you can't do anything. It's the upstream who must force them to extend a proper license to you.
However there was also a case recently where the FSF argued that this interpretation / handling of the situation is against the spirit of GPL and I think they won, so... Yeah, it's just unclear. Which is normal for legal texts (IMHO intentionally, but I'm not here to rag on lawyers, so I'll leave it at that).
While I agree with your view (at least when it comes to firmware, especially given that hardware that doesn't require a firmware upload on boot generally just has the very same proprietary firmware on a built-in memory, so the only difference is that you don't get to even touch the software running on it), the point of this project is to remove non-libre components from coreboot/libreboot.
It doesn't differentiate itself from upstream in any other way, so if it fails to do the one thing it was made to do, then that's in fact a newsworthy fact.
I do not know of any such dongle, but I'd like to ask you a question if you don't mind: are you looking for a dongle with open-source firmware, or would a dongle that has its (proprietary) firmware stored in some onboard memory be acceptable?
The second option wouldn't require you to install any proprietary firmware on your computer, but you'd still rely on the proprietary firmware for the device to run. And it might also exist, unlike a dongle with FOSS firmware.
I know this isn't Reddit, but r/peopleliveincities... When 90% of desktop users use Windows, it's going to both be the most targeted by malware developers and have the highest chance of being operated by someone who doesn't understand enough about computers to recognize that the shiny calculator app that just popped up after visiting a very legit Nigerian prince's crowdfunding page probably shouldn't need admin access.
And speaking of user error, I'm willing to bet that basic security practices like using full disk encryption, SecureBoot, some MAC layer (provided by antivirus on Windows, AppArmor/SELinux on Linux) and regularly applying security updates are way more common over in the Windows land - if I was in a situation where there was one completely randomly selected Windows PC and one also completely randomly selected Linux PC, and my life depended on being able to gain access to either of them (some kind of really messed up Saw trap? idk), I would definitely bet my life on the Linux one being misconfigured.
Don't get me wrong, Linux can make for a very secure and private OS, but most installs most definitely cannot be described as such - just look at the popularity of random unverified PPAs on Ubuntu derivatives or AUR packages on Arch.
A reasonable build of the kernel optimized for virtualization won't take more than a few tens of megabytes of RAM (and it will have support for memory ballooning, so the virtualized kernel will give the memory it doesn't need back to the host), and the userspace will need to be separate anyway due to how different Android is to normal Linux distros.
Containers are nice when you want to run dozens of separate services on the same server or want to get the benefits of infrastructure as code, but in this case they would provide minimal benefits at the cost of having no way of loading any kernel modules not built into whatever ancient kernel version your SoC manufacturer decided you have to use on your phone. Also, container escape vulnerabilities are still a bit more common than full VM escape, so this is also good for security on top of being more useful.
Idk, ~five years doesn't seem like a long time with regards to CPUs. I had my Ryzen 2200G from summer 2018 to this January, and I would consider my purchase of a cheap Ryzen 3600 to be a bit premature (and it was definitely an impulse purchase). And it also means that the CPU sticker became outdated pretty soon.