edit: one of my previous attempt: https://news.ycombinator.com/item?id=35229211
I actually have made it extensible, with closely coupled source of rules and domains; but then I lost it Edge forgot all my userscripts :(
And most user scripts are so long a typical user won't be able to spot a couple of malicious lines amongst 10k lines of minified webpacked libraries.
Exactly!
That's why you should use 3 lines for it instead, that are
- inspectable
- not updateable by the Chinese/Russians
- written by you anywayFor example, if you use FreeTube with SponsorBlock to improve your privacy and block ads, in fact you are sending to Cloudflare 100% of your YouTube watch history, and to SponsorBlock ("sponsor.ajay.io").
With Piped instances it's even worse, essentially escaping Google's tracking just to give our data to random strangers.
If you are worried, just run a second Chrome session with NordVPN and uBlock Origin in a loose jurisdiction and browse YouTube unlogged.
It's easy, simple, and you have the benefits of an audited platform and that reasonably legally confirm they don't store logs unless the court forced them: "we never log their activity unless ordered by a court never log their activity unless ordered by a court", but for that, the court has to find you as a user, which can be very complicated in practice.
So much better than random strangers.
I feel like I’m on YouTube already.
It’s not like they are free of criticism either.
Tho I probably should've demonstrated first that it is possible, before advocating for it. The script I linked indeed only works for one website. Multiple websites with multiple rules, each with a list of instances (that often go offline for a time, so it is worth keeping them around, and make switching easy) indeed complicates it a bit.
"having to code all the rules" is not that hard, in most cases you can just pass the whole URL, and the instance accepts it.
Advantages: you don't get unwanted redirects from services, and you don't get unwanted redirects to instances. (Even tho the information about the instances will likely be concentrated at libredirect github issues. Chances are that some random person on the internet who has paranoid activities as a hobby will look into the instances, so you don't have to.)
- - -
I don't use many redirects. Nowadays I use exactly 0. But if I needed a redirect for example to xcancel, I would use my user-script as I had done it in the past before I lost it. I definitely wouldn't install a browser extension for it.
In all cases that also involves actually finding the URLs, then there are non-most cases where a slice wouldn't do it.
> Nowadays I use exactly 0
Exactly. If you ignore actual uses everything becomes trivial
If you actually did this you would know that it works for all of a week or two before YouTube stops letting you watch videos until you login.
NordVPN only sees that you connect to YouTube, they do not see the pages or videos that you are looking at, and from the perspective of YouTube, they only see requests from a very popular VPN where are millions of users.
If you use the "privacy" instances, these "privacy" websites and Cloudflare knows precisely which videos you are watching.
Also IIRC for youtube, alternative frontends don't tend to rely on someone else's endpoints.
I was considering reformulating it, in $CURRENTYEAR there is always someone that claims that using Russian or Chinese as a synonym for 'enemy' is Russo- and Shinophobic. I've decided against it this time.
[0] : https://en.wikipedia.org/wiki/Synecdoche#Part_referring_to_w...
I'd much rather send random tidbits of information, that are nearly useless in isolation, to strangers than to the central tracking corporation
In the end, there is no way to reveal what information you're interested in when retrieving data, short of retrieving a ton of data and doing the filtering client-side, which is also an option with these third parties if you so desire