Windows is therefore going to be used by a wide variety of systems, including those with low grade hardware parts. So it either has to be part of windows quality control to test these low end systems or it should make it very clear about the system requirements.
Either way, the problem was patched, which makes it a problem on windows side.
If this was a widespread problem I'd agree it should've been caught during testing and mitigated, but it doesn't sound like that's the case ("We looked around and could not find other reports resembling such situations"). As far as I'm aware it's just one Twitter thread reporting this, not all/many low end systems. Only so much Windows can do if it's just a faulty batch of SSDs, or even an incorrectly installed part in the user's setup.
> Either way, the problem was patched, which makes it a problem on windows side.
Has it been? I can't see Microsoft even verifying that the issue exists yet.
Linux still has tons of driver issues and is a massive time vampire with its "type magic spells into a terminal for hours to get something done" GUI.
Mac is way more expensive and doesn't run a ton of apps available on Windows.
cough Linux, BSDs, Solaris clones cough
If you have an old car, you shouldn't press the gas pedal. /s
Sure, Microsoft can issue a SW workaround to mitigate this issue as they did for Intel Skylake CPU issues, but that doesn't change the fact that the core issue is faulty designed HW.
Why don't SSD manufacturers test their shit before shoveling it out the door to consumers?
Most of them do, not that it stops people from buying Temu PCs anyways.
Damn, i hate to defend Mickey$oft here but, if your hardware breaks under work loads, better buy a better one. Windows is not able to use all cores on a CPU properly, so if you have a workload problem, it is your hardware.
SSD makers should make better products.
Microsoft should do enough testing so it doesn't push updates that trigger data-losing hardware flaws.
Microsoft should be more careful, but also such blatantly garbage hardware never should’ve shipped. So I’d say hold MS responsible, but also push back against manufactured e-waste like that.
What does "more careful" even mean? For instance, if your crappy SSD has some bug that causes data corruption if there are more than 16 writes in flight at a time, and microsoft pushes an update that causes it to do writes more aggressively, should it really be microsoft's fault that your crappy SSD broke? Should microsoft keep throttle windows IO at 2000s HDD throguhput on the off chance that some IDE drive bricks itself when it encounters SSD level workloads?
It seems increasingly these problems (including the NVIDIA one) are only cropping up on 24H2 machines which is why I've decided to stick with 23H2 until it's out of its support window. Unfortunately that's coming up pretty soon.
[1]https://www.neowin.net/news/nvidia-fixes-windows-11-24h2-dri...
see sibling comment for something similar happening on zfs. linux is also filled with workarounds for crappy SSDs, so "SSDs breaking due to software" isn't exactly exclusively a windows problem. It just gets more attention because of the bigger user base.
>This suggests that this kind of strain on the hardware is unnecessary given the task, which then paints a picture of negligent sloppiness in engineering and QA culture at MS.
"straining hardware" shouldn't be a thing for PC components. If it can't handle a given workload, it should throttle/queue the requests, not brick itself. I can grant some leeway for long term wear (eg. heat damage or electromigration), but that's clearly not what's happening here. Can you imagine a network card that bricks itself if you send too many packets, and you have to baby it with how many packets you send in any given amount of time?
I don't think Windows has the most hardware compatibility issues.