18 ms·

tl;dr can someone summarize for non-German readers? Why?

by jitl 9y ago

lightedman 9y ago
They claim it is to offer you more diverse search results in the face of centralization.
- floatingatoll 9y ago
  I can validate this claim, having been in some of those discussions at work.
toyg 9y ago
Google Translate is pretty good on this page. Basically, they invested in this Cliqz company that "provides recommendations and other information, such as news, weather news, or the world of sports, while typing into the address line". Being a recommendation engine, it needs to be fed user data, so they are automatically shipping and turning on user tracking in 1% of Firefox downloads from German servers. For a brand built on privacy and freedom, it's shocking - but it's the same sort of idiocy that we saw with the Pocket thing a few years ago. Clearly nobody learnt the lesson - or maybe they did, this time they are sneaking things through slowly and communicating it on Friday afternoon. This is pure filth from Mozilla.
482794793792894 9y ago
Well, first of all, I wouldn't call it "tracking". They're going to be aggregating the data and specifically promise to not create profiles on individual users. Secondly, I wouldn't necessarily call it "third-party" either, since Mozilla owns parts of the Cliqz GmbH, so will at least know what's going on inside the company and can easily pull the plug, if Cliqz should actually start to store user data. Generally speaking, though, Cliqz is a content recommendation engine, trying to compete with web search engines, so sort of hoping to take a bit of world domination (and money) out of Google's hands. The actual recommendations will be generated locally without sending data to Cliqz, but they do need a data set that tells them what users are usually looking for based on the browsing history that they have.
- mthoms 9y ago
  >The actual recommendations will be generated locally without sending data to Cliqz If this is true, that's great news. They should mention it in the post IMHO.
  - floatingatoll 9y ago
    I suspect it would make little difference. Either users are okay with server-side data collection (with differential privacy etc), or they are hostile to all data collection full stop. The in-between nuance of “results generated locally” does nothing to salve the anguish of this post’s comments. I wish it would.
- Vinnl 9y ago
  > They're going to be aggregating the data and specifically promise to not create profiles on individual users. A promise isn't good enough. They say so themselves either, which is why they've described here how they try to prevent themselves from even being able to create profiles, so you don't have to trust them: https://gist.github.com/solso/423a1104a9e3c1e3b8d7c9ca14e885e5 https://gist.github.com/solso/423a1104a9e3c1e3b8d7c9ca14e885... That said, a cursory reading does make it seem like they wouldn't be completely unable too, but I haven't dived into it too extensively yet.
floatingatoll 9y ago
Search engine lock-in is a serious and real problem to us. Search engines aren’t effective vs. long-tail sites. A search engine that takes into account recent browsing is vastly more capable of surfacing relevant results. Google has done this for years under the head on “Personalized Search Results”, but continues to force users into Google Accounts to derive a benefit, complete with the horrendous privacy policies that comes with. We’re trying to solve this without violating privacy. We encourage and work with the TOR project, which builds ESR with all server-side functions disabled. If you require inaction, request denied.
- userbinator 9y ago
  A search engine that takes into account recent browsing is vastly more capable of surfacing relevant results No, it's vastly more capable of "surfacing" sites you've already visited, putting you into a filter bubble. If you really care about privacy, switching the default search engine to DuckDuckGo would probably make a better statement of that.
  - 1024core 9y ago
    > No, it's vastly more capable of "surfacing" sites you've already visited, putting you into a filter bubble. Not necessarily. Here's an example: say the 3 most visited sites by a user are NYT, LAT and the Atlantic. Even if this user has never visited The NewYorker, it's a good bet that she would find it of interest.

tl;dr can someone summarize for non-German readers? Why?

by jitl 9y ago

tl;dr can someone summarize for non-German readers? Why?

lightedman 9y ago

They claim it is to offer you more diverse search results in the face of centralization.

floatingatoll 9y ago
I can validate this claim, having been in some of those discussions at work.

toyg 9y ago

Google Translate is pretty good on this page. Basically, they invested in this Cliqz company that "provides recommendations and other information, such as news, weather news, or the world of sports, while typing into the address line". Being a recommendation engine, it needs to be fed user data, so they are automatically shipping and turning on user tracking in 1% of Firefox downloads from German servers. For a brand built on privacy and freedom, it's shocking - but it's the same sort of idiocy that we saw with the Pocket thing a few years ago. Clearly nobody learnt the lesson - or maybe they did, this time they are sneaking things through slowly and communicating it on Friday afternoon. This is pure filth from Mozilla.

482794793792894 9y ago

Well, first of all, I wouldn't call it "tracking". They're going to be aggregating the data and specifically promise to not create profiles on individual users. Secondly, I wouldn't necessarily call it "third-party" either, since Mozilla owns parts of the Cliqz GmbH, so will at least know what's going on inside the company and can easily pull the plug, if Cliqz should actually start to store user data. Generally speaking, though, Cliqz is a content recommendation engine, trying to compete with web search engines, so sort of hoping to take a bit of world domination (and money) out of Google's hands. The actual recommendations will be generated locally without sending data to Cliqz, but they do need a data set that tells them what users are usually looking for based on the browsing history that they have.

mthoms 9y ago
>The actual recommendations will be generated locally without sending data to Cliqz If this is true, that's great news. They should mention it in the post IMHO.
- floatingatoll 9y ago
  I suspect it would make little difference. Either users are okay with server-side data collection (with differential privacy etc), or they are hostile to all data collection full stop. The in-between nuance of “results generated locally” does nothing to salve the anguish of this post’s comments. I wish it would.
Vinnl 9y ago
> They're going to be aggregating the data and specifically promise to not create profiles on individual users. A promise isn't good enough. They say so themselves either, which is why they've described here how they try to prevent themselves from even being able to create profiles, so you don't have to trust them: https://gist.github.com/solso/423a1104a9e3c1e3b8d7c9ca14e885e5 https://gist.github.com/solso/423a1104a9e3c1e3b8d7c9ca14e885... That said, a cursory reading does make it seem like they wouldn't be completely unable too, but I haven't dived into it too extensively yet.

floatingatoll 9y ago

Search engine lock-in is a serious and real problem to us. Search engines aren’t effective vs. long-tail sites. A search engine that takes into account recent browsing is vastly more capable of surfacing relevant results. Google has done this for years under the head on “Personalized Search Results”, but continues to force users into Google Accounts to derive a benefit, complete with the horrendous privacy policies that comes with. We’re trying to solve this without violating privacy. We encourage and work with the TOR project, which builds ESR with all server-side functions disabled. If you require inaction, request denied.

userbinator 9y ago
A search engine that takes into account recent browsing is vastly more capable of surfacing relevant results No, it's vastly more capable of "surfacing" sites you've already visited, putting you into a filter bubble. If you really care about privacy, switching the default search engine to DuckDuckGo would probably make a better statement of that.
- 1024core 9y ago
  > No, it's vastly more capable of "surfacing" sites you've already visited, putting you into a filter bubble. Not necessarily. Here's an example: say the 3 most visited sites by a user are NYT, LAT and the Atlantic. Even if this user has never visited The NewYorker, it's a good bet that she would find it of interest.