Google's new Privacy Sandbox

[Owen Lucas]

So, instead of Tom's Hardware advertisers using cookies to track the browsing habits of your IP, including your browser type and the like (and of course your ability to block this with disabling 3rd party cookies or using UBlock and similar), you'll get ads chosen by Google based on what Chrome sees you doing.

I can see a lot of private ad companies having issues with this, calling Google a monopoly by shutting out another part of the market. I obviously don't know all of the moving parts here but continuing to limit advertisers data like this makes targeting harder. I know iPhones rotate mac addresses for privacy reasons and use various other tricks to make individual device ID tracking harder. Maybe the ad companies know they are slimy and will just use another tool in their data mining toolbox.

I mean that's good for the consumer though right?

 

[BaronHK]

I can see a lot of private ad companies having issues with this, calling Google a monopoly by shutting out another part of the market. I obviously don't know all of the moving parts here but continuing to limit advertisers data like this makes targeting harder. I know iPhones rotate mac addresses for privacy reasons and use various other tricks to make individual device ID tracking harder. Maybe the ad companies know they are slimy and will just use another tool in their data mining toolbox.

I mean that's good for the consumer though right?

Most of Apple's privacy campaign is just marketing.

I've found so many bugs in Webkit that any time you load a Web page on your iPhone, there's about a gazillion bugs for zero click malware. Which is exactly what you don't want to hear about something whose only job is literally to contain untrustworthy content.

There were even some where the debugger gives up and crashes and eventually I gave up, but I found nearly four dozen bugs of varying severity in a single month and it wasn't even Apple that fixed them, it was Igalia. It took Apple seven months to get those out to their users, probably while State actors or even Pegasus was using some of them.

Their code isn't fantastic. And it sprawls. About the only parts that may be extremely well written are whatever's still in there that came from KDE in the 2000s.

Almost all of the bugs and poor design that I've ever seen in software was something Apple threw over the wall, like ALAC, or messed with, like AIFF. They put JPEG2000 in their browser when nobody else did and now there are things that return jp2s when you send a Webkit UA, and they break if you don't have it, so now we need a really buggy image library that hasn't been looked at much for security issues, pointed straight at the Web.

In WebkitGTK, it helps that the bubblewrap sandbox contains the Webkit Web processes, but it would still be better if we could drop the library.

As to "Google Privacy Sandbox", never ever call it that. It's basically another tool for browser fingerprinting, although there are so many now that I doubt it's all that interesting by itself. Apple is responsible for another big problem called Canvas. But there's a million bits of information a user's Web browser tells servers that you can't suppress.

 

[Owen Lucas]

It took Apple seven months to get those out to their users, probably while State actors or even Pegasus was using some of them.

By design?

Most of Apple's privacy campaign is just marketing.

Thanks for the deep dive into this. A little sad it's smoke and mirrors!

 

[alarmguy]

I’ll take Apple any day over any of the other browsers.
Sandboxed, tracker blocking, and IP address hiding.
Their privacy controls. I certainly think is at least is good as the best out there.

Not too long ago, there was only one company Facebook was going to sue because of the privacy controls and that was Apple. Call me crazy but that is certainly a statement and it’s documented initially cost them well over $1 billion in advertising revenue. I certainly put more credibility in that unless somebody can come up with something better.

Certainly would like to know who is better. It would help everybody else.

 

[BaronHK]

I’ll take Apple any day over any of the other browsers.
Sandboxed, tracker blocking, and IP address hiding.
Their privacy controls. I certainly think is at least is good as the best out there.

Not too long ago, there was only one company Facebook was going to sue because of the privacy controls and that was Apple. Call me crazy but that is certainly a statement and it’s documented initially cost them well over $1 billion in advertising revenue. I certainly put more credibility in that unless somebody can come up with something better.

Certainly would like to know who is better. It would help everybody else.

Facebook could likely slip the iOS sandbox. But they'd end up with a PR debacle and potentially Apple throwing them out of the app store when it was discovered, and maybe suing them. (which would be delightful...I don't use Facebook because I know what it is.) Apple's review process is hardly bulletproof. They screamed when Epic left sleeper code in there that made it through the review process because it wouldn't run until after they figured the review would be over and it would be rolled out to users.

It's much easier to attack Apple in court, with lawyers and judges, so if Facebook does something I'd absolutely expect it to be in court.

Apple claims that installing software packages directly could cause the iPhone to become a malware sewer. F-Droid for Android isn't a malware sewer, and it's hard to have that happen when something is open source because anyone can view the source code and propose fixes and alert others, and revert "features" that they don't want.

But if you pirate Android apps, certainly your phone can become a malware sewer. That's the main reason not to even try it.

We've never had a serious malware problem on Linux because the software distributions tend to be open source software packages and the package manager verifies that the package hasn't been tampered with since it was put on the server, and the connection to the server with the packages tends to be secured. I can probably count on one hand over a quarter of a century where some malicious code got uploaded somewhere and ended up on some computers, like the gnome look incident back in like 2007 I think. It's never really from the distribution's own sources, iirc. It's so rare I don't even think about it. I just stick to trusted sources of packages.

Android phones can be a lot worse than iPhones, or a lot better. It depends on what you install on them. The Play Store apps almost always have too many trackers in them, even forced into the packages if the developer didn't put them there, because Google does.

That's why the Play Store is collecting dust on my phone and I use programs from F-Droid, including a corrected version of Firefox (Fennec F-Droid). No banking apps. No government apps. NOTHING from a car insurance company.

At one point, I read a thing that said that Allstate was spying on people through GasBuddy and a weather app, and selling their driving habits to LexisNexis to go on their CLUE report. Even for non-Allstate customers.

Doesn't help to not take the "discount" for putting spyware on your phone if you just install the spyware somewhere else.

People who write proprietary software tend to always be looking for a way to make money off of "free" (of charge). People who write open source software tend to do it because it's their baby. There might be a donation box or something.

Even on my Chromebook, I don't use Google's cloud. I have a 4 TB external SSD, lots of internal storage, SD cards. LibreOffice. I know what "cloud storage" is. It's data leaks, right? It's warrantless searches. It's Google scanning your files to see what's in them to sell ads.

At one point, a creep with administrator access who worked for Google was inappropriately using his access to try to prey on minors.

I know someone who worked there until the late 2010s. He said there were still many engineers at Google that had that kind of access, including your Google Drive, but the Access Control system logs who is using that.

I've also seen how Google complies with warrants and National Security Letters. It almost never fights them and in some cases it's prevented by law from telling the user it even happened. The user will find out when they have a shiny new set of handcuffs, or maybe never find out if the investigation closes out and they didn't find anything.

At the very least, Google Drive is an expensive way to back things up that are then only available with an internet connection, and only at the speed of that connection. Expensive compared to the one time cost of just storing your documents and files locally and paying for that storage once.

 

[BaronHK]

By design?


Thanks for the deep dive into this. A little sad it's smoke and mirrors!

I don't know if it's by design or not. There's a lot of bad programming practices at Apple. It's possible that they strategically delayed while State Actors got the heads up, but I'd just chalk it up to lazy and slovenly practices.

The Shellshock bug in bash was fixed in bash upstream very fast and rolled out to users. Apple didn't like the GPLv3 license and had stuck to a version of bash that was over a decade old from 2007, and now they use zsh I think. It took them a lot longer to fix that bug because they would have to either grab the latest version of bash that fixed it for them and had all sorts of cool new features that programmers would like, such as associative arrays, or write their own patch.

You can get the latest bash on your Mac from something called Homebrew and set it as your default user shell if you like, but how many users even know or care? It's not a serious UNIX system. It's been a long time since they marketed Mac OS as something geared towards techies. Ken Thompson recently got rid of his Mac and moved to Linux on a Raspberry Pi.

I work with audio. It's still easy to use my Chromebook Plus in Crostini as a digital audio workstation using Linux software. I even have Wine, so some of the software is for Windows. In the mid 2000s I came across some AIFF files, and found out that Apple, in the change from PowerPC to x86, had used a total hack to flip the byte order (endianness) of AIFF. They flagged it, incorrectly, as "AIFF-C" when it was not compressed, and then they put in an uncompressed audio stream inside the file with the byte order flipped to LSB first. When the more obvious solution was to write some code in the playback side to anticipate that the stream could be big or little endian.

I mean, I looked at that and I was like "What the heck am I even seeing here?" Because first I thought the file was corrupt because that's what my software told me. Now the software knows about Apple's byte-flipped AIFFs of course, but they didn't announce to anyone outside the company that they would do this.

Their crusade against GPLv3 is that if some of it happens to touch DRM, then they would have to give the user the means to unlock the DRM and admit that it is not an effective technical measure to prevent copying. However, I can't see how anyone would make the case that bash is an integral component of DRM. It also doesn't touch their verified boot sequence governed by the T2 chip.

As soon as you install Crostini on a Chromebook, you have the current version of bash and they load it in Termina. I've set things up so that I can quickly load the fish shell though because the scripting syntax is more comfortable for someone that has a background in older BASIC and batch scripting, or even Rexx.

 

[OVERKILL]

I don't know if it's by design or not. There's a lot of bad programming practices at Apple. It's possible that they strategically delayed while State Actors got the heads up, but I'd just chalk it up to lazy and slovenly practices.

The Shellshock bug in bash was fixed in bash upstream very fast and rolled out to users. Apple didn't like the GPLv3 license and had stuck to a version of bash that was over a decade old from 2007, and now they use zsh I think. It took them a lot longer to fix that bug because they would have to either grab the latest version of bash that fixed it for them and had all sorts of cool new features that programmers would like, such as associative arrays, or write their own patch.

You can get the latest bash on your Mac from something called Homebrew and set it as your default user shell if you like, but how many users even know or care? It's not a serious UNIX system. It's been a long time since they marketed Mac OS as something geared towards techies. Ken Thompson recently got rid of his Mac and moved to Linux on a Raspberry Pi.

I work with audio. It's still easy to use my Chromebook Plus in Crostini as a digital audio workstation using Linux software. I even have Wine, so some of the software is for Windows. In the mid 2000s I came across some AIFF files, and found out that Apple, in the change from PowerPC to x86, had used a total hack to flip the byte order (endianness) of AIFF. They flagged it, incorrectly, as "AIFF-C" when it was not compressed, and then they put in an uncompressed audio stream inside the file with the byte order flipped to LSB first. When the more obvious solution was to write some code in the playback side to anticipate that the stream could be big or little endian.

I mean, I looked at that and I was like "What the heck am I even seeing here?" Because first I thought the file was corrupt because that's what my software told me. Now the software knows about Apple's byte-flipped AIFFs of course, but they didn't announce to anyone outside the company that they would do this.

Their crusade against GPLv3 is that if some of it happens to touch DRM, then they would have to give the user the means to unlock the DRM and admit that it is not an effective technical measure to prevent copying. However, I can't see how anyone would make the case that bash is an integral component of DRM. It also doesn't touch their verified boot sequence governed by the T2 chip.

As soon as you install Crostini on a Chromebook, you have the current version of bash and they load it in Termina. I've set things up so that I can quickly load the fish shell though because the scripting syntax is more comfortable for someone that has a background in older BASIC and batch scripting, or even Rexx.

You sound like a guy with a whole lot of coding experience! You work for one of the big houses for a few decades?

 

[BaronHK]

You sound like a guy with a whole lot of coding experience! You work for one of the big houses for a few decades?

Most of the things I write are to make my own life easier. Without going into too much detail I'm not allowed to talk it involves data recovery.

The overall details of the process are not embargoed but obviously you don't want competitors to know too much.

To make my life easier I switched to the Chromebook because the OS is good about managing itself in the background while I spend most of my time in Crostini and the separation and security work done by Google keeps things isolated so I don't have to worry much about data being stolen or ransomware attacks. You really can't run executable code in the Chrome OS layer if you want to because it's not meant to be used that way.

I have Borealis Steam, but that's about it as far as generalized executables (within Steam). The rest is in the Crostini VM.