Ubuntu has this package installed by default:
network-manager-config-connectivity-ubuntu

It's only purpose is to provide settings for NetworkManager to send requests to connectivity-check.ubuntu.com , and based on the result (AFAIK) detect redirection by captive portals and open an ISP's page (think public WiFi, or hotel rooms, where you need to authorize to access the net).

Well, connectivity-check.ubuntu.com is hosted on Google cloud (you can check that by running:

dig connectivity-check.ubuntu.com
whois [the IP from previous query]

), so by default Ubuntu sends requests to a Google cloud page.
I don't say Google counts daily active Ubuntu users (because many of those have the same IP), or that Google actively logs and analyzes that data. But some of you guys may not like that behavior.

So what's the fix?

Purge the package

sudo apt purge network-manager-config-connectivity-ubuntu

If you do need a captive portal detection, create your own config file to query some HTTP (not HTTPS) page of your choice, in the example below I have a Debian page used for the same purpose. Use your favorite text editor to create and edit /etc/NetworkManager/conf.d/90-connectivity-custom.conf :

[connectivity]
uri=http://network-test.debian.org/nm

Restart NetworkManager

sudo systemctl restart NetworkManager

If you run an Ubuntu derivative, please report if you have network-manager-config-connectivity-ubuntu installed in the comments.

In light of the seemingly increasing rate of data breaches and privacy violations in general, I've decided to take some steps further regarding my online presence.

Among other things, I decided to switch all my online accounts to custom domain email addresses, so I grabbed two domain names (with WhoisGuard enabled): one for use with stuff related to my real identity (think @firstlast.com), and the other for all else (think @randomword.com). Then, I changed the email address of each one of my existing online accounts, taking advantage of the catch-all feature. To make things short, it goes like this:

Accounts not related to my real identity:

• tildes.net.187462@randomword.com -> tildes.net
• reddit.com.178334@randomword.com -> reddit.com
• ...

Accounts related to my real identity:

• amazon.com.113908@firstlast.com -> amazon.com
• bankofamerica.com.175512@firstlast.com -> bankofamerica.com
• ...

As you might have guessed, the 6 digits ending the local part of email addresses are meant to be randomly generated, in order to mitigate easy guesses by spammers due to catch-all (though I've also created a specific sieve filter to mark incoming emails with "unknown" recipient as spam).

Before you ask, I don't intend to start a discussion about threat modelling here. I just want—as anyone who is not a complete tech-illiterate—to have a reasonable weapon against spam caused by recurrent data breaches, so that if an email address is leaked, I can toss it and replace it with a new one without much effort.

Also, I value owning my email addresses, in the sense that if I decide to change email provider in the future, I won't have to change my addresses too as a consequence. For communicating with real humans (e.g., my doctor), I could use a non catch-all address like first@firstlast.com.

I wonder what do you think of this approach... Is it overkill? Do you see any major concern from a privacy or security standpoint? Are you doing something similar and are happy with it? I would very much like to hear your experiences with email, especially about the approach you settled with.