Blows my mind that there aren’t common modpacks for Skyrim. Last time I tried getting into it I spent probably a week getting everything together… then launched the game, played a couple of hours, then got distracted by life.
Never went back to it because I didn’t want to go through the exercise of maintaining it.
It also happens from image preview caches created by explorer. I see this get hung up for some time with SMB shares and the like that have images.
Something is still interacting with the photos on your device, and although it may be mundane… do you really know every single service and process running on your device, all the time? Could we ever know? Just takes one dependency for one thing you installed to be bogus… it could even be from a rootkit installed in bios that installs whatever software on bootup so even if you wipe your system it’s there, ever monitoring, ever feeding your data away.
There’s just no user friendly tooling on windows that’s built-in which would ever pick this stuff up. AV doesn’t know what is desired or undesired behavior when it comes to stuff like this either. Sure, it won’t send up stuff protected by UAC from a non-admin request… but thumb drives, CF cards, SD cards etc all have no restrictions.
I have power toys… well on my windows dual boot anyway. Do normal users? Probably not.
I’ve used unlocker tools in the past as well, but this is not something provided in the OS and as such not something a typical user is even aware of, let alone has available.
You’ve never, ever plugged in a thumbdrive and then used the “safely remove hardware” tray to try and eject… only to receive an error that says “unable to remove, in use” or similar?
I guess i’ve never seen it happen when I was using a thumb drive to image a machine well over a decade ago.
There are a few people on the internet who seem to have reported this though:
https://www.reddit.com/r/WindowsHelp/comments/pnbfcf/this_device_is_currently_in_use_when_trying_to/
https://learn.microsoft.com/en-us/answers/questions/4212072/cant-eject-external-hard-drive
There are thousands of results of this. I have encountered this dozens of times personally across dozens of different systems.
How do you know that the latest version of a piece of software that auto updated on your computer doesn’t have anything snooping on your hard drive and uploading whatever it wants somewhere?
Think about it. How would you know? Let’s assume it’s not using a bug, exploit or vulnerability.
Windows hasn’t been my main os for a while but I’m fairly certain you can mount/unmount drives without rebooting.
I work in IT for a living. Sometimes something keeps your drive locked. Windows does not confess. I wasn’t talking about linux user experience because most people don’t use linux like we do.
Imagine if you were a malicious actor and you wanted a copy of all photos someone plugged into a computer that were not things like browser cache, just good honest to god OC.
All you have to do is listen on drive letters D, E, F, G and when one is plugged in with a DCIM directory… silently upload the data contents to a server over the internet when a drive is detected with that subdirectory.
Have you ever wondered why you couldn’t eject a drive without rebooting? It’s not like it’s going to tell you what process is keeping it locked… Encryption wouldn’t even matter, because you’re gonna need to decrypt/unlock it to access it, and windows doesn’t care what service or application is trying to access it, it is glad to allow any kind of file action without even admin rights.
Anywho, actor has your photo, AI trivially builds facial recognition models, pulls in timestamps, geolocation metadata, camera metadata… and now those photos you never intended to upload anywhere are in a database of PII that will be shared to god-knows-who.
Cool. Pop_OS is working quite well over here.
It’s never been a better time to convert to Linux. The best part is that if one distro turns to shit you can just move over to another.
I’ve seen this but I don’t really want the docker part.
I think it could be phenomenal for some kind of beefy VDI implementation for low demanding games or some kind of monster server with multiple GPUs, but it just feels wrong for an individual who wants to remotely stream their desktop on demand and has no plans on having others share the host.
Maybe i’m overthinking it, or haven’t thought it through enough - but my gut says this has more drawbacks than i’m realizing.
Why? for a feature that doesn’t exist consistently or comfortably on windows?
non-arm-Windows client to Windows host you can use something like parsec where you can adjust resolution on the fly. You can install a fallback virtual display for when the display is off from right inside the app in a single click + UAC prompt. I really don’t like parsec though because I know the enshittification will come, and I don’t really trust them to be secure or to not abuse their backdoor accesses.
This project allows you to create virtual displays that are persistent, and you can configure them so that when the primary monitor is off the backup one is enabled in windows, just by using the default windows display manager options. You can change the resolution freely… because this is using the same vd driver parsec created https://github.com/nomi-san/parsec-vdd - this works pretty well overall with Sunshine
Ultimately though, Sunshine and Parsec are the only two things i’m aware of with great low latency and high fidelity remote capabilities, aside from niche implementations like what the PS5 has. If something like Xorg had similar quality and latency parity i’d be interested, but i’m under the impression everything is like old school vnc or rdp where it is functional when necessary but not very pleasant overall.
It cannot generate a virtual display. It only uses attached displays which by default are real powered on monitors.
I’ve gotten around this on windows with parsec and a virtual display adapter that someone keeps updated on GitHub which can spawn backup displays if none are present, but I find still sometimes fails to spawn them. Parsec is fairly reliable at spawning them when the windows solution fails but it’s not perfect either.
A hack job virtual display on Linux will be more difficult to work with. It’s going to eat my desktop and be fairly hidden.
Dummy plugs exist, but I specifically don’t want to put dummy plugs in all my remote hosts. Seems like an unga bunga solution to something which should be software.
Something with VNC or simply ssh with some scripting could be the workaround I use to get back in when a virtual display fails to work as expected, but I am lazy and want something effectively bulletproof.
…and just to be clear, this is a multiplatform problem. There’s a single mediocre ‘easy’ option in windows land and a very tinkery option in linux land.
Doesn’t seem like any OS has caught up to the idea of fast streaming desktops quite yet. I’m convinced it’s the future of computing though. Way better than old VDI options from days of yore.
Switching to linux has been the best decision i’ve made all year.
Just wish there was a good one-click-setup virtual display option for Sunshine that “just works.” It’s my white whale of features.
Sure, but trying and failing is the only way we’ve ever accomplished anything.
With the rate of CVEs scored at 9+ that come across all the stuff I manage at work I would not agree.
It would be foolish to simply stop patching this giant pile of obsolete insecure dependencies they are calling windows.
The fact that in lieu of paying them they simply want you to log in with an account they control should speak volumes to the privacy implications.
Must be a reason why they don’t want us to see the code. Maybe it’s whatever they’re changing to attempt to build a walled garden to keep the app protection money rolling in.
What purpose do electric door locks serve? Other than being more prone to failure, more expensive, and dangerous?
An oligarch’s fancy?
I’m sure in product meetings it’s been brought up that it’s a dumb thing and they could save money and make the cars safer by not having them, then the oligarch speaks up.
If they can identify the user’s computer, they can cross reference against all the other tracking metadata available. Since they know the browser down to the individual, they also know the predominant IP of said browser and can link your actions from before and after the vpn/proxy account was created.
As long as javascript is running they can track your mouse movement, which is similar to people’s gait when they walk. it’s unique, it’s identifiable. You can probably fuck with them by using a trackball on one profile, but once they link that trackball movement to your profile with other metadatas then the cat is out of the bag and you’re permanently known to Reddit and whoever it sells it’s data to.
As much as people will claim “just disable javascript!” - you’ll find that you practically cannot use the internet without it… and having JS disabled makes your fingerprint even more unique as few disable it carte blanche.
Basically, once they have a shadow profile of who you are it’s just a matter of time for them to link any account created to it. I suspect almost everybody’s shadow profile is quite complete.