Gregory's Server
@denschub In this environment, intentionally giving the advertisers *yet more* data can only harm privacy.
|
5 comments
@mcc @denschub calling something "privacy-preserving" because it prevents individual re-identification is misleading, because most privacy issues are group level and most privacy harms are perpetrated by classifying someone as a member of a group DP is great for browser telemetry when you want to know that someone's browser crashed doing some specific thing but you actively don't want to know who it was—but it's not a good fit when the recipient of the message has a motivation to discriminate  @mcc @denschub the ‘yet more’ is what gets me. this whole thing seems to rely on some pie-in-the-sky belief that ad tech is going to stop doing all the invasive shit they’re doing now in favor of this opaque, trust-us-it’s-not-invasive-despite-the-invasive-launch ‘feature.’ with no reason to believe this is true, ff can only be adding to their data stash. and yes, dennis, i’ve read your opinions on these points here & on reddit. they have only further eroded my trust in mozilla. it’s damning.  @brhfl @mcc @denschub we want to be clear, we were a party to the early strategy discussions on this stuff in our time at Google (yes, that was five years ago; yes, it's taken quite a while to gain the traction the company wanted) and it's a desperate play BY the ad industry. the public should not go along with it. |
@denschub It doesn't matter if I can, at this moment, find a deanonymization attack on Firefox PPA. It matters if marketing firms or nation-states can find one. If they do find one, they won't be sending you a nice email telling you about it.
Given all this I feel "privacy-preserving" is a deeply dishonest descriptor, and differential privacy—made of advanced math, something most consumers are not good at—creates in the consumers' mind a perception of certainty that you cannot actually provide.