It's good to see Krebs ending his piece with a conversation starter on the costs/benefits of using Facebook, and whether people are changing their usage or deleting their accounts. I wonder when this discussion will expand from tech sites to the general press?
Years ago Krebs did something similar whenever there was a new hole in the Java interpreter. He regularly advocated that people remove Java support from their web browsers, and that they consider removing it entirely from their computers if they don't use any Java apps.
In a related comment found below the original article, "MrB" writes how Facebook still tracks people without accounts. This is why I block Facebook's BGP prefixes in my router, their domains in my DNS resolver, and key hostnames in /etc/hosts. With this combination, Facebook is unlikely to be able to track me even when I'm in between setups. For example it should work in cases where I temporarily fail to re-populate /etc/hosts while doing testing, or set up a new computer (still using the same router), etc. And since their IP blocks and domains are relatively stable, it doesn't take much babysitting on my part.
It's pretty minimal, at the moment only connect.facebook.net, connect.facebook.com, pixel.facebook.com, and apparently facebookinc.122.2o7.net. If I had to rely solely on it I'd add a lot of hosts and in particular known ones under fbcdn.{net,com}. In contrast, DNS and IP blocking is extensive.
The implication was that these groups probably mostly consist of scammers scamming scammers and may act to deter the kind of cybercrime which actually affects innocent people.
Years ago Krebs did something similar whenever there was a new hole in the Java interpreter. He regularly advocated that people remove Java support from their web browsers, and that they consider removing it entirely from their computers if they don't use any Java apps.
In a related comment found below the original article, "MrB" writes how Facebook still tracks people without accounts. This is why I block Facebook's BGP prefixes in my router, their domains in my DNS resolver, and key hostnames in /etc/hosts. With this combination, Facebook is unlikely to be able to track me even when I'm in between setups. For example it should work in cases where I temporarily fail to re-populate /etc/hosts while doing testing, or set up a new computer (still using the same router), etc. And since their IP blocks and domains are relatively stable, it doesn't take much babysitting on my part.