> Microsoft provided the FBI with the history of IP addresses tied to that specific GDID.
This article, and most articles about this, doesn't explain where FBI got that GDID from. Ok, Microsoft has a list of IP addresses that has been used by a computer with a certain GDID, but FBI needs to get the GDID in the first place, and then try to bind that to a person.
I found another article that explains the process a bit better:
> Stokes got caught because he used the same Windows device for everything, and the GDID stitched all of it back together after the fact.
> Scattered Spider members phoned the jewelry retailer’s IT help desk from Google Voice numbers, posed as locked out employees, and talked support staff into resetting three accounts, two with administrator privileges. From there they installed a tunneling tool called ngrok to get past the retailer’s network defenses, moved roughly 77 gigabytes of data to Amazon cloud storage using ngrok [...]
> Investigators later subpoenaed ngrok and found the account used in the attack had been created on May 12, 2025, at 19:21 UTC from a VPN proxy IP address run by Tzulo, a hosting provider. The IP was a dead end. VPN proxies do that. But the GDID is built different.
> Microsoft’s records showed that at that exact same minute, a Windows device carrying GDID g:6755467234350028 had visited the ngrok signup page. Three hours later, the same GDID visited the retailer’s own website, through the same Tzulo proxy address used to set up the ngrok account. It gave the FBI a device, that don’t rotate the way VPN exit nodes do.
> Microsoft has a list of IP addresses that has been used by a computer with a certain GDID, but FBI needs to get the GDID in the first place
What they did was the opposite: ask Microsoft for GDIDs used by attacker-associated IPs within several 24-hour time periods during which attack-related activity took place. Windows pings Microsoft regularly with the GDID, establishing links between your GDID and any IP addresses you use. The IP logs from Microsoft and the VPS provider showed at least 10 instances where a single VPN IP accessed the attacker's VPS and also pinged Microsoft with at least one GDID within a 24-hour period. They found a constant GDID that all instances shared. This seems to have been the most damning GDID-related evidence in the DOJ complaint [1] and yet it wasn't mentioned in the article you linked (or any other articles about this I've seen pop up on HN). It includes the diagram from the complaint (page 18) that outlines this, but devoid of context. The ngrok stuff that the article focuses on was just the cherry on top and was discussed later in the complaint.
What also becomes clear when you read the complaint is that the GDID was just one piece of the puzzle and that they had plenty of other evidence. Attacker-associated IPs were used to access the suspect's Apple, Snapchat, and Facebook accounts, at least one of which was his actual residential IP, not a VPN IP. Once they had revealed the identity of the person who owned these accounts, they were able to all-but-confirm that this was in fact the attacker.
What remains unclear even after reading the complaint is how they were so sure that the GDID they obtained visited specific websites, but honestly, at that point, they were already drowning in evidence, so I don't know if it matters that much. It could be as simple as "he was signed into Edge with his Microsoft account and had sync enabled".
Wait, so it would have, ironically, been safer to allow microsoft telemetry to bypass the VPN entirely and remain associated only with their home network, because it's the phone home to microsoft tunneled through the VPN that tied together all their IP addresses to a single microsoft account. The GDID itself is almost a red herring, as it could have been a session id or username or something only long lived enough to appear comming from several ip addresses to windows update request?
If they'd done their VPN/dirty work in a linux VM/container, with the VPN running in that, they'd have been fine.
Clean-them would have had their GDID on their normal ISP IP or equivalent, and Dirty-them would have had everything through the VPN from their no-telemetry dirty-host.
This seems like something that should be easy to confirm, but I haven't seen anyone do it. Do they keep a database of every website visit all Edge users make?
Speculation: if you turn on browser sync so you can have your open tabs and history on different computers, that stuff will be on MS' cloud with your consent, with whatever E2EE they have or claim to have.
But it would be very foolish for a blackhat to turn on browser sync...
I got banned from all computers in my secondary school and from taking the IT GCSE; final school exam in the UK.
Using the school news paper shared account, I copied the bullies coursework in to the public share, made changes to the coursework on my own user account and then copied it back to their work folder using the school news papers account.
MS Office keeps an "Last edited by" field so I got caught. If I had used the school newspaper account to make the edit I wouldn't of been caught. Rookie error. I too discovered a DCOM in the Windows 98 help file that revealed all the hidden shares on the network which scared the school. This was 2004 and I was 15.
Oh and the time, I bought a BB gun off someone at school. I was a librarian in the schools library and showed it off someone of the older grade year. The head librarian confiscated it and reported it to the headmaster. The guy then gave me a copy of Q3A and I stuck with violence the video-game way.
They didn’t. They went to ngrok and asked for all the data at the point of signup. They then looked to find the any of that data at the second site. In this case they had two identical data points - the GDID and the IP address.
What do you mean? The FBI could know that this VPN ip visited ngrok and then the retailer website, but how would it know that that ip was associated with the specific GDID unless microsoft was tracking (timestamp,website,gdid) tuples?
Data Description for Browsing History data type
Microsoft browser data subtype: Information about Address bar and Search box performance on the device
* Text typed in Address bar and Search box
* Service response time
* Autocompleted text, if there was an autocomplete
* Navigation suggestions provided based on local history and favorites
* Browser ID
* URLs (may include search terms)
* Page title and notification text
They claim it was an ngrok account that was used to host an endpoint used in the compromise, tied to a microsoft account / gdid that was passed when ngrok software was downloaded from the "microsoft store".
This criminal mastermind got caught because he did everything but sign his name to the crimes while holding two pieces of government identification in presence of a notary.
The FBI did the bare minimum in terms of old-fashioned detective work, and correlated evidence from various sources.
The obsession with GDID is a complete nothing-burger and I'm tired of seeing it on the front page every other day.
Windows 10 LTSC IOT will be the last Windows I ever use.
As I grow older, I simply do not have the patience or the will to do all these workarounds and tweaks to my OS to turn it from a piece of barely-working corporate spyware into something that I can call a productive tool.
It also seems like interacting with the OS is on its way out anyway, as most people essentially interact with computers via the browser (essentially a different sandboxed OS altogether), whether on desktop or mobile device.
Interesting, generally Microsoft bypasses the hosts file name resolution for various MSFT domains. Curious that these were not included (if it works, which I assume the mitigation does).
FWIW, YogaDNS stops these bypasses if you tick a settings box ("Block plain DNS over TCP from System Resolver"). Or use similar DNS forcing techniques against port 53.
Windows falls back to the normal resolver in that case.
I think it's likely that Microsoft is running a process to correlate "new" GDIDs to old ones, ex:
"Oh look, this one has almost all the serial numbers of components and attached-devices as that other one, it's probably the same computer with a fresh install, let's make a note of that..."
They do not need this, they require you to create a Microsoft account to use your own computer (currently without a phone number, passport and selfie but that will probably change in the future).
(Aware/assuming you're being sarcastic.) I've always felt that statement reveals a serious lack of imagination, or at least a failure to apply it to the question.
It's not one homogeneous people reacting to different OSes. It's different people, and they react differently when somebody else's OS does something vs when their OS does something. Windows users don't care or feel locked in because it's the only OS they've known and they depend on it. Linux users expect this type of behavior from Windows, but they can't do more than switching themselves which they already have. Were Linux to do something similar, you remove the offending piece of software.
But you don't have a service uploading it along with the URLs you visit, which is the creepy part. The ID itself can be useful for your own administration. And if such a service were to appear, it could easily be removed. Everything in Linux is optional, especially as long as you stick to open source.
How do we know? Does anyone check for things like this?
Many, many apps read /etc/machine-id if you do a quick github search.
Apps may have been silently correlating our activity for years without us knowing.
We know DHCP, EFI, GNOME, popularity-contest and many other apps already use it. There are countless ways it could be used already that are hard to detect.
Basically all source is open. Eyes are plenty. Somebody would raise a stink. People can also notice things while monitoring their network via a number of tools like Wireshark, etc. All it takes is one person to notice and share their findings.
It also helps that "Linux" isn't a monolith. One person's installation can be very different from another in terms of the software used. If one piece of software collects, that collection isn't as valuable as Microsoft's because the userbase is much smaller.
It's not perfect, but it's a hell of a lot better. A lot more resistant to abuse. Security in a lot of contexts tends to be relative like that. One's house isn't impenetrable; it's just better than another, in part because the neighborhood is better.
By way of bugs (unintended effects in code), who knows how it compares to proprietary codebases. Those also depend on the particular companies.
However, by way of intentional code that goes against the interests of the user, which would be considered security issues with FOSS but not with proprietary codebases where the interests of the company come first, things should be better with FOSS.
The latter should also be more noticeable than the former. It's possible some of the former may not have run at all, neither accidentally nor intentionally. A bug can just stay dormant without ever surfacing in practice. The latter is there to be used.
The benefit of FOSS with respect to eyes is that those eyes are better aligned with the interests of users (since it's the users looking) than the eyes on closed-source code.
It should be noted that the "correct" usage of /etc/machine-id is that you use it in a way that doesn't allow cross-correlation between different applications by using a HMAC of the machine-id with an application-specific UUID instead of using it directly. systemd-id128 has a command line flag to do this for you.
Whether everyone does that correctly, that's a different question. On the other hand, while developers should be more careful about how they use identifiers like this, if an application on your machine wants to track you they don't need /etc/machine-id.
Mobile's came at the cost of composability between programs and the imposition of policy, which are against the Unix Philosophy. I wouldn't make that trade. Mobile's a sad state of affairs, inferior to desktop in many ways.
You can most likely get better security than mobile's, you just need to e.g. learn to write your own SELinux policies, etc. Facilities are there; they just have a learning curve.
How do you write SELinux policies to allow reading only certain files in /proc, where process IDs are not known ahead? I ended up writing my own FUSE-based /proc emulation. The facilities are there, but it feels like writing your own OS.
What kind of use-case do you have for that? I suppose whatever it is, you could also e.g. write a privileged service that checks those files with whatever security policy you need. Your client wouldn't have direct access to /proc.
Another option may be to set up a container or PID namespace and give your tool direct access to that /proc.
The /proc contains too many unnecessary information, which can be used for fingerprinting or helping an attack. Run `ls /proc` and find yourself surprised. For example, why do applications need to know the kernel command line? What for? Why do they need the list of major and minor device numbers? List of filesystems? Network configuration?
So I want to follow the principle of minimal privileges and only grant access to files needed for running a program. Sadly many programs cannot run without /proc. For example, poorly coded Apple's Grand Central Dispatch library crashes the application (for example, Telegram) if it cannot enumerate the information about threads or processes. It needs this information to calculate how many additional worker threads need to be created, and if it cannot calculate the number, it terminates the application for reasons I do not understand.
There is also a catch that the program can create a new unprivileged user namespace and mount /proc there thus bypassing my daemon completely. Anyone can create a user namespace nowadays.
> Another option may be to set up a container or PID namespace and give your tool direct access to that /proc.
/proc contains information not only about processes, but a lot of extra information.
On most distributions I've seen, SELinux is configured to automatically label /proc files with labels based on fairly granular roles, so that setup already exists.
If you want to completely remove the global information in /proc you can do so by mounting it with subset=pid (and extra points for hidepid=4). This was added to improve container runtime security but unfortunately a lot of programs still depend on reading global procfs files so we can't enable it by default.
As for why all this information exists in /proc, procfs has historically been a bit of a dumping ground. There is an argument for "everything is a file" and so on, but having had to deal with all sorts of insane bugs related to procfs I'm less sympathetic to that argument than I used to be.
> There is also a catch that the program can create a new unprivileged user namespace and mount /proc there thus bypassing my daemon completely. Anyone can create a user namespace nowadays.
There actually are some restrictions, if the process doesn't have access to an unmasked procfs mounting /proc will fail, though a recent patch[1] finally made it so that subset=pid works in this case. (And most sandboxes disallow unprivileged CLONE_NEWUSER with seccomp.)
I also wanted to limit access to files/directories inside /proc/PID, or provide fake data. The /proc/PID also contains too many data.
Also, I might want to provide some fake data, for example, if app relies on reading /proc/cmdline or /proc/cpuinfo. So the app should be able to read the files, but not the real data.
> And most sandboxes disallow unprivileged CLONE_NEWUSER with seccomp
I ma worried that some applications (like Chrome, Electron-based apps) might not work without user namespaces. GTK uses "glycin" library, it launches subprocesses for handling images in a bwrap-based sandbox, and it broke inside my sandbox, so I had to do quick fixes for it.
No, sandboxing in Linux is not easy. Just look at "man capabilities" or "man user_namespaces" and see how many rules and exceptions from rules are there. You need to understand it if you write a sandbox, but it is so complicated. And obviously AI cannot be trusted with such a responsible task.
By the way I just read about OpenAI model breaking into Hugging Face and it used "cat /proc/self/mountinfo" to collect information about third-party sandbox it wanted to break from [1]. So my intuition that these files should not be accessible to the sandboxed program, was correct.
It sounded like your problem was restricting access to some /proc/<pid>/ dirs and not others without knowing the PIDs ahead of time:
> How do you [...] allow reading only certain files in /proc, where process IDs are not known ahead?
Such things like /proc/cmdline should be even easier. You can just set regular DAC permissions and ownership on such files. You can just set ACLs on such files.
> There is also a catch that the program can create a new unprivileged user namespace and mount /proc there thus bypassing my daemon completely.
You don't get the ability to mount just because you created an unprivileged user namespace.
> For example, why do applications need to know the kernel command line? What for? Why do they need the list of major and minor device numbers? List of filesystems? Network configuration?
Because you might ask them to get that info. If you wouldn't have a use for that, you prevent that using one of a number of security mechanisms, including writing your own SELinux policy where you don't whitelist that access, setting DAC permissions and ownership as such, ACLs as such, etc.
Applications compose and are heavily configurable, and they include such things like language interpreters. You may want to write a script that uses any of that info for whatever. You may want e.g. your window manager to show that info on a statusbar periodically. That info can also be useful for conky or htop, etc.
Vim is a text editor, but you can e.g. have it make a call with ssh to get info from wherever and put that info into the buffer. You can say that a text editor doesn't have a need to read network configurations, but maybe by composing with ssh it does, and it does something useful with it.
I am not sure if I can set ACL or ownership on /proc files, it is not a disk-based filesystem.
> You don't get the ability to mount just because you created an unprivileged user namespace.
Please consult "man user_namespaces", section "Effect of capabilities within a user namespace". It clearly says that one can mount /proc inside a mount namespace governed by a user namespace without being root. The reason why you weren't aware is probably because the rules related to namespaces and capabilities are extremely complicated.
> Vim is a text editor, but you can e.g. have it make a call with ssh to get info from wherever and put that info into the buffer.
I definitely do not want proprietary applications running ssh on my system without me knowing.
It's hard to do a 1-to-1 comparison but if the telemetry also used a masked form of the machine-id with an application ID unique for the telemetry service then you wouldn't be able to correlate them directly.
To be clear, I think this whole story stinks but the issue here is not that OS installations have unique identifiers (machines have lots of unique identifiers -- hardware MAC addresses, peripheral composition and device IDs, screen resolution) -- it's that Windows has an auto-enabled spyware^Wtelemetry system that shares enough data for them to be able to provide that kind of data to law enforcement. This would be a privacy nightmare even in the absence of a unique OS installation identifier. Heck, they could even take regular screenshots of your screen but that would be too obvious[1].
Indeed. One can regenerate it on each shutdown / reboot. The process is a little different depending on whether one has systemd or not. It's a dbus thing but systemd ingests it and there is a specific process around updating that in systemd.
There is also the NetworkID in Firefox about:networking#networkid
Another trackable piece of information on most systems is the creation time of / which just about any application can query unless it is properly isolated. This can be turned into a short unique hash. There are hacky ways to change the Birth time on unmounted filesystems using debugfs which may result in corruption. Some overlay filesystems do not support Birth time but that is not going to help most people unless the general populous expect all applications to be isolated in name spaces and overlay filesystems but this would have to be an expected pattern across all applications universally.
stat / | grep irth
Birth: 2023-04-17 20:27:01.000000000 +0000
stat / | grep irth | md5sum
8b5e849954373f8f3c2a625847e4e858
The trick is to combine it all, which is what MS is doing. For example your MS account stores any GDIDs that you've logged in with.
If you use a different identifier to track changes in a GDID/machine-id/etc., that means you can continue tracking using mainly just the new machine-id, but you should always keep trying to correlate it with other things in case it changes.
I agree, tracking data via unique identifiers is evil incarnate, unless it's Google or VC-backed adtech doing it, because how else will they make money having architected their entire businesses around it.
I think the difference is that, on Windows, there are background services that constantly ping Microsoft with the device ID. A device ID on its own is not really harmful if it's not exposed to the internet.
Right, the danger here isn't stable unique data itself--there's already plenty of that--but the OS "telemetry" which steals [0] it and reports to Big-Brother along with too much other betraying information. Every site you visit, every program you run, the serial numbers of all your hardware, etc. Even if the GDID were totally absent, it would still be a correlate-able privacy nightmare.
Ultimately there's no informed consent here: The average consumer is disbelieving and surprised if you tell them what kinds of stuff Microsoft has/can put into a dossier. Nobody thinks: "Ah, Edge on a fresh Windows install, I'm glad Microsoft knows every site I visit, and can tell I'm a friend with someone because we use the same bluetooth speaker."
[0] It seems wrong to use the verb "leaks" when it's so obviously intentional.
I mean, there's a ton of unique identifiers on machines already tied to hardware and disks, but that must be a systemd thing since my Devuan machine does not have it.
But given there's no cloud accounts on linux I would imagine it's trivially changed just like a NIC's MAC
Also, seems unlikely it would be used for any single-signon with cloud services.
You're right. I have no idea what it's for, but I'm guessing for distinguishing between deployment images. I'm certainly not going to get too concerned about it.
The issue with the microsoft account was not unique IDs - tons of unique things on a machine.... it was a unique thing tied to an account and shipped to remote places.
There are _some_ cloud accounts for linux such as Ubuntu One, which I would fear could have similar identification capabilities if ever forced by to do so by a 3-letter agency
Any sort of online service you're logged into is going to have your IP and account correlated with a timestamp. Dropbox, GDrive, and even your e-mail provider. While a device identifier might be more useful it's not like the lack a device ID will keep you from being tracked.
I won't be surprised if MS patches workarounds quickly and won't shed a word on the whole situation. Or worse, pulls the "this helps fighting evil hackers" reasoning.
But it is regenerated by reinstalling. So I suspect that its less fingerprint based and more of an in disk IDs thing. Which goes back to if you clear the right things while the system is not running you might be able to change it
>Every finding here was reproduced on a real Windows 11 Pro VM (build 26200). Nothing is theoretical. See docs/technical-writeup.md for the evidence, tagged by confidence level.
That's sounds nice and all, but everything about it, from "Nothing is theoretical" to "evidence, tagged by confidence level" goes out the window when you find claude as one of the commit authors, and the content is clearly copy pasted output from claude with very little editing. Worse yet, one of the sources he cites is also clearly AI output.
I'm not even against the use of AI here. I would rather see it clearly say either "this is what claude found after I told it to investigate" or "yes this is generated by claude, but I independently verified each of the points myself".
If you care about privacy, you simply cannot use Windows. Microsoft has made it clear over the last decade (if not longer) that they have a vested interest in compromising your ability to use software without exploiting and monetizing data about your usage.
Microsoft is a post-privacy corporation. Eventually, we really have to stop acting so shocked about this sort of thing from them.
What Google was "caught" doing was exactly what it was doing in non-incognito mode already. Anyone expecting the websites you visit to know you're in private mode and to not log data is just lacking basic tech knowledge.
It is silly that Google had to explicitly state that in the disclaimer.
It's reasonably detectible. And it wasn't about "websites" logging you in general, it was about Google in particular logging you, the same company that was making the misleading privacy promises. They deserved the slap over it. And it required a bit more than basic tech knowledge to understand what was misleading, it was a real problem.
"Now you can browse privately, and other people who use this device won’t see your activity. However, downloads and bookmarks will be saved."
Can you tell me how the original Incognito mode disclaimer was misleading?
Even if it is detectable it shouldn't be detected, why would I want websites to act differently when I'm in private mode?
And if it did block Google's services from capturing the same data as they capture in non-private mode it would act as a sort of ad blocker, why should that be a function of private mode?
"privately" was off which is why it says "more privately" now. The combination of "privately" and "chrome won't save your history" isn't a good enough way to explain that the rest of Google is actively countering your attempt at privacy.
> why should that be a function of private mode?
Because it's the same company that was promising the privacy.
>"privately" was off which is why it says "more privately" now
The actually significant change is the addition of "This won’t change how data is collected by websites you visit and the services they use, including Google."
Which was obvious to anyone with basic tech knowledge already.
>Because it's the same company that was promising the privacy.
But WHICH privacy? It never promised that websites won't know who you are, you can still login into websites, you're still in the same IP. Should it automatically disable all cookies and route you through seven proxies so you can be truly "private"?
The part about not collecting data. The kind of stuff chrome.exe is still doing.
I don't even agree it was obvious to anyone with basic tech knowledge that google.com would still super aggressively track you. Third parties yes, but not google itself. That was a policy decision they made. (Any 'how could they know' is post-hoc justification far more than real technical problem.)
This article, and most articles about this, doesn't explain where FBI got that GDID from. Ok, Microsoft has a list of IP addresses that has been used by a computer with a certain GDID, but FBI needs to get the GDID in the first place, and then try to bind that to a person.
I found another article that explains the process a bit better:
> Stokes got caught because he used the same Windows device for everything, and the GDID stitched all of it back together after the fact.
> Scattered Spider members phoned the jewelry retailer’s IT help desk from Google Voice numbers, posed as locked out employees, and talked support staff into resetting three accounts, two with administrator privileges. From there they installed a tunneling tool called ngrok to get past the retailer’s network defenses, moved roughly 77 gigabytes of data to Amazon cloud storage using ngrok [...]
> Investigators later subpoenaed ngrok and found the account used in the attack had been created on May 12, 2025, at 19:21 UTC from a VPN proxy IP address run by Tzulo, a hosting provider. The IP was a dead end. VPN proxies do that. But the GDID is built different.
> Microsoft’s records showed that at that exact same minute, a Windows device carrying GDID g:6755467234350028 had visited the ngrok signup page. Three hours later, the same GDID visited the retailer’s own website, through the same Tzulo proxy address used to set up the ngrok account. It gave the FBI a device, that don’t rotate the way VPN exit nodes do.
https://www.windowslatest.com/2026/07/10/you-cant-fully-disa...
Although this doesn't explain where Microsoft got that traffic data from. How do Microsoft know which sites a computer visit?
reply