Fair enough

I’ve heard of being blocked from a specific service like Xbox Live but never the whole of Microsoft. How would people get Windows updates? It’s crazy.

Mainly because energy and data centers are both expensive and companies want to use as little as possible of both - especially on the energy side. OpenAI isn’t exactly profitable. There is a reason companies like Microsoft release smaller models like Phi-2 that can be run on individual devices rather than data centers.

How did that happen?

Yep and somehow people who don’t know better are up voting him. Not surprising for this platform.

Yeah it’s not always that simple. You haven’t been around long enough to see the stuff that can go wrong with installing Windows. For example I recently had Windows refuse to see both SSDs in a machine. All because of something called Intel VMD. Took me a handful of attempts before I found the problem.

When Windows installs work they are fairly simple if long, but when they don’t work oh boy.

The unplugging of internet to get a local account?

Also they disabled that for Windows Home.

Some Lemmy users are actually just wankers. I would like it if you all stopped. It’s especially great when I have people like you who probably aren’t even experienced in tech.

This actually happened to me before recently and all it took is one firmware setting. So frustrating.

Actually no. It’s not Mint’s decision whether to start the install USB with UEFI or BIOS. It actually depends on what the firmware chose to start and how the install medium is formatted. Some install media is only setup for BIOS booting, some for only UEFI, and some can do both. If the firmware detects the medium as supporting both then it should choose UEFI first but this depends on what settings you have in the firmware, and if you choose an option at a boot menu as boot menus allow you to override the default. When it comes to actually installing the OS most sane installation software will look at how it booted and install that way. So if it detects it was starting with UEFI it will configure the install to be UEFI, same if it was started with BIOS it will install as BIOS. How does it know? UEFI variables are one way. They can normally only be accessed if the system was started with UEFI.

If you truly wipe a drive you wipe the partition table as well. You say the table is outside the file system formatting, and this is sort of true, but they are both just data on the disk. Disk don’t care where the partition table ends and the file system begins. In fact you don’t even need a partition table at all. Unlike some other systems Linux will let you put a file system straight on the disk, the whole disk, with no partition table in sight. It’s not recommended mind you, because it will freak Windows out if it sees it. Windows will see it as a blank disk and not so helpfully offer to format the thing. When I say format a disk, I mean the whole thing, partition table and all. It’s also not possible to make a partition tableless disk bootable in UEFI. In BIOS it’s possible though as BIOS doesn’t read partition tables. It just needs a boot sector and that’s it.

Also if you’re trying to change a disk from MBR to GPT, and you don’t care about data, you shouldn’t be converting it. You should be formatting/wiping the whole thing and making a new partition table. Which is normally what it offers to do if you tell it to erase everything and install it.

Edit: Getting down voted for actually knowing how computers work and bothering to explain it. Shock horror.

UEFI won’t boot from MBR drives unless it’s in BIOS compatibility mode. What format the drive is in isn’t determined by a firmware setting, though it can affect the boot process. I don’t think you actually understand what you are talking about here. The easiest way to install OSes both Windows and Linux is by wiping the drive, which would have solved this issue. Dual boot on single drive configurations normally have issues and will always be more complicated. It’s better to use two drives where possible in most cases. I suggest you read up on BIOS vs UEFI and how partition tables work if you want to do a complex setup like that.

Mint is known for having older kernels and therefore not supporting the latest hardware. They have a different edition for newer computers called Linux Mint Edge edition. Something Arch derived like CachyOS or another distro using recent kernels will always have the best support for bleeding edge hardware. The CachyOS installer is also pretty friendly, though maybe not as much as Mint.

This isn’t true. Try Linux Mint or Ubuntu, their installers are much better. Those installers used by Fedora, RedHat, and even SUSE can be a bit weird.

They specifically say unbloated Windows as well which while it’s not as difficult as they make out is still somewhat annoying.

I’ve recently had a Windows installer fail to see my NVMe drives until I changed some random UEFI setting because it was missing a driver. Linux could see it just fine, as could Hirens boot.

It was an improvement but still not great. Ideally they would have kept the Windows 7 interface with maybe some upgrades like virtual desktops, then continued with all the under the hood improvements.

You can also rollback changes on arch using snapper.

How unstable is the unstable branch exactly?

Why doesn’t lemmy at least have pinned comments and flairs? Seems like a serious omission to me.

ChromeOS literally is Linux so obviously it can do everything that Linux can. It is effectively a SUPERSET of Linux

This is dumb. I was experimenting with unofficial ChromeOS builds since before the first consumer Chromebook. I’ve also used an actual production Chromebook as well. I’ve even used the distro ChromeOS is derived from which is called Gentoo.

If this was actually a good idea why aren’t some businesses doing it already? Linux servers are everywhere yet I have never heard of one running ChromeOS. Google who make ChromeOS don’t use ChromeOS for servers, they use conventional Linux distros like Debian with their own software running on top such as Borg or Kubernetes.

First I am going to assume you mean a Linux distro (e.g. Debian, Ubuntu) rather than just Linux, as all distributions are a superset of just Linux, as Linux is only the kernel.

It’s also not in anyway a “superset” of a conventional Linux distribution. It doesn’t even have a package manager without having to use a VM. You can’t install a different browser without using a VM. ChromeOS capabilities have improved a lot since I have used it, by supporting VMs at all and by allowing Android apps. That still doesn’t compare to a real Linux server solution. On a Linux machine I can natively run containers, not just VMs. I can install apps natively. I can configure my own security systems and sandboxing. I can even run Windows apps and games with Wine and Proton, android apps with Waydroid. I also get a much newer kernel, older Chromebooks didn’t get kernel updates throughout their entire life. Maybe that’s changes but who knows.

Another big thing Linux servers have is advanced file systems like BTRFS and ZFS, that have advanced RAID like functionality with automatic data integrity protection using checksums, snapshots built in, and other fancy features like transparent file compression. Does ChromeOS even support regular software RAID?

ChromeOS just isn’t designed for servers, it never has been. You’re trying to use a car to haul a lorry load. It’s not that cars are bad, they just aren’t designed to do that. You could use Chrome OS for employee workstations, at least some of them anyway, but not for servers.

Go and use a real Linux distro or a real Linux server and get back to me.

Now you are being ridiculous. We are talking about code than runs in the kernel but is not part of any official kernel module including device drivers.

This is false. That module is signed by Microsoft. That means they tested it themselves. To load a module that isn’t signed on Windows requires serious tinkering and is something no business would do. I have actually done those steps myself, so I have direct personal experience here. If you had watched that video I sent you, which is by a retired Microsoft Engineer, you would know about this.

Dude you are the person who thinks that the Cloudstrike code running in the kernel that is neither part of the kernel nor part of any official kernel device driver code is somehow equivelent to the actual kernel. You are also the person who made the completely nonsensical claim that ChromeOS Linux “can’t do nearly the same things” of Linux. So GTFO with the snarky shit, because I’m the one wondering how you can be so confused about basic stuff.

For one you apparently don’t listen. I said it’s a kernel driver/module not a core kernel component. It does however still run in kernel mode (ring 0 on x86) and has access to everything any other thing running in kernel mode has access to including the NT kernel itself. It doesn’tq matter from a permissions perspective if it’s a driver/module or a core kernel component, the CPU protections don’t distinguish between the two by design. What does make a difference is when and how it’s loaded into kernel space/kernel mode. That’s why safe mode works, as it just doesn’t load that component.

I am not the one getting basic things wrong here. When I used ChromeOS originally it had maybe half the capabilities it has now, but even now it just isn’t as capable as a conventional Linux distro. Saying it can run VMs means nothing because so can regular Linux, in fact you can run ChromeOS in a VM, or Windows, or FreeBSD all on a Linux machine. Even at the same time if you really want.

I’ve been using the internet since 1983. How bout you? LMFAO.

Using and understanding are not the same. I have a Masters in CS, soon to be starting a PhD in Cyber Security. You meanwhile apparently have no experience or qualifications worth telling me about. I actually use Linux systems daily and run my own Linux based servers. Specifically I use Proxmox since that’s an actual server solution.

Because you didn’t do enough research. My Chrome OS comes with an outer VM, an inner VM, the heavily locked down user mode, and the kernel mode. My Linux programs run in the outer VM by default. I would have to turn on developer mode just to get to the actual user mode.

I did plenty of research. It says the VMs run inside of a container, but a container is not a Virtual Machine. It is technically a form of virtualization, so you’re half right. I would take the time to read up on how containers actually work. They share the same kernel as the host machine, and don’t emulate hardware the way virtualization solutions do.

Also none of this is part of the default setup, which doesn’t include any Linux VMs, just the locked down user space. You actually can do Virtual Machines inside Virtual Machines on Linux. It is called nested virtualization. My hypervisor Proxmox actually supports that as standard. You can also do containers inside virtual machines and virtual machines inside containers. The former I actually use as part of my own server setup. The point being none of this is unique or special in ChromeOS.

No you don’t have any evidence at all. You were specifically speaking about Chrome OS using VMs inside VMs. That article doesn’t mention VMs once.

I never once disputed that it was more secure than an average Windows installation, because frankly that’s obvious. What you don’t seem to understand is what can be done to lock down modern Windows and Linux systems.

Yes it was a ridiculous system to have kernel mode code on the filesystem. Even if a bad pointer didn’t crash the system a hacker could have put in their own code. And yes such a terrible security system would have affected Linux too. But with ChromeOS, the system is already secure. No need to use a terrible security system like Cloudstrike in the first place.

For one the same article you are referencing talks about using anti-malware on ChromeOS as Chrome OS isn’t malware proof. Though I don’t think it’s possible for Chrome OS to break so badly from one of these products.

ChromeOS though isn’t actually suitable for running servers like Windows and Linux are. It can’t do nearly the same number of things. It’s a bit like comparing a knife to a safety razor. One is safer for shaving sure, however the other one can be used for cooking, hunting, wood work, etc.

Second all kernel mode code lives on the filesystem. How did you think it worked? On Linux and Windows the kernel itself needs storing somewhere, as do the modules.

You keep showing me again and again that you don’t understand the world of computers and modern IT infrastructure. Do you even have any qualifications or work experience in IT?

Edit: I actually did some research myself. ChromeOS can use a Virtual Machine to run Linux software, but not in it’s default configuration. There is also none of this VMs inside VMs stuff you were talking about. If you want to see virtualization really put to work look at a modern server setup or something like Qubes OS.

https://chromeos.dev/en/linux/linux-on-chromeos-faq#can-i-access-files-when-the-container-isnt-running

Says person who thinks Visual studio is the best IDE

Automatically reverting the kernel wouldn’t do anything as the kernel nor the module Crowdstrike uses were updated. Rather the file the module reads was updated and replaced with a corrupted version that causes the module to crash when it tries to read it.

There is a great video explaining the basics of what happened here: https://youtu.be/wAzEJxOo1ts?si=_agkbdBHJnhQmbdP

Microsoft already have a mechanism to disable problematic modules on next boot. Problem being that Crowdstrike registers itself as an essential driver, as they don’t want the system to boot without it for security reasons.

You keep saying Chrome OS uses VMs for security. Unless something has seriously changed since I last read up on their security mechanisms they don’t. Maybe something has changed. Do you have any evidence? If not you’re just talking out of your arse.

Any of the JetBrains suite.

I’m guessing Crowdstrike issues a lot more Windows updates than Linux updates?

Not really. Linux is used for critical servers everywhere. No reason to update it less often.

It’s not that Linux can’t have security problems. I still remember the very first internet virus in 1987 that traveled thru Unix machines. But Windows is the worst OS for critical systems precisely because it is the most common OS. Anything is better than windows. Linux, MacOS, or even an old IBM mainframe OS and those awful tn3270 terminals. Also, Chrome OS in particular has VMs instead of other VMs. It really is designed to be much more secure than Windows

This isn’t a hacking attempt. It’s not a security breach. None of the “Windows is more common” stuff is actually valid in this case. The fact it’s not actually true is even more funny. When it comes to servers and smartphones, and the total number of devices in general, Linux outnumbers Windows. Linux isn’t actually niche in the slightest, only purists running Arch or Ubuntu think that because they ignore any Linux they don’t like, like Android.

You also don’t understand anything about ChromeOS security either. They don’t use VMs for system security. Early Chromebooks actually had virtualization disabled! Sandboxing and virtualization aren’t the same thing. The reason it’s secure is largely because it can’t do anything, it uses an a/b root system, and it has secure boot by default. It’s not that fancy anymore.

Chrome the web browser also has sandboxing on Windows, and modern Windows uses secure boot. Edge and Chrome have the same foundation btw. What it’s missing is the immutable a/b root system, and the fact Windows allows running arbitrary executables when ChromeOS doesn’t. There are actually tools for making Windows immutable, and with group policy or things like S mode you can restrict who can run what executables. Meaning with the right settings it’s almost as secure as ChromeOS. Even more funny Windows actually does use virtualization based security. So you have that backwards too.

Stop talking about shit you don’t understand and learn about it instead.