Most of the interesting data mining happens at the client, not the server.
- I can not put tracking cookies on other people's clients. Twitter puts tracking cookies, can collect location data, can get your contact list...
- If any instance starts manipulating messages from its users to add tracking redirections, people would immediately lose trust over it and drop it. With twitter, there is absolutely nothing you can do and people treat t.co as an inevitable necessity.
- Even if someone went on to analyze the data and tried to use in the messages to give people personalized ads in some custom client, what would be the incentive for people to stick to it? If the publisher wants to have a business model where they want to share ad revenue with the users, they might as well go with something like Brave.
Comments
What’s to stop them from data mining ever server they federate with?
Like, especially, a bad actor could run several large instances with no public relationship and so all kinds of correlation.
Most of the interesting data mining happens at the client, not the server.
- I can not put tracking cookies on other people's clients. Twitter puts tracking cookies, can collect location data, can get your contact list...
- If any instance starts manipulating messages from its users to add tracking redirections, people would immediately lose trust over it and drop it. With twitter, there is absolutely nothing you can do and people treat t.co as an inevitable necessity.
- Even if someone went on to analyze the data and tried to use in the messages to give people personalized ads in some custom client, what would be the incentive for people to stick to it? If the publisher wants to have a business model where they want to share ad revenue with the users, they might as well go with something like Brave.
I’m thinking of stuff much more insidious than ad targeting - things like de-anonymizing
Then you do threat modeling in the same way as you would with email.