Slashdot Mirror
Facebook Announces Social Search Tools
Today at a press conference in California, Mark Zuckerberg announced a big new feature from Facebook: Graph Search. It's a set of tools designed to quickly bring together social information involving "people, photos, places, and interests" in response to a user's query. Zuckerberg was quick to point out that they aren't indexing the web, and thus aren't challenging Google. However, it will use the vast volumes of data already stored on Facebook to answer questions like "What kinds of movies do my friends like?" and "Who are friends of friends that are single in San Francisco?" Addressing the obvious privacy concerns, the company said it wouldn't allow users to search content that wasn't already shared with them (or already public). The searched data does, however, include location data, if it's been shared — you can search by places your friends have been. Significantly, the official site also mentions that Graph Search will help you meet new people, something Facebook hasn't really highlighted until now. Graph Search is being rolled out as a limited beta, with only a few thousand participants. In the coming months, they'll open it to more users and continue working on mobile and non-English versions.
Additional levels of automated stalking!!!
When I've wanted to know what movies my friends like, I'll probably have already talked to them about it.
On Facebook, though, I've got "friends" who are basically just people I shared some period of time and space with - e.g. high school classmates. I don't really care what movies they like, unless they're members of the tiny minority with whom I've kept contact over the decades.
BTW this is the exact same logic that made me immediately turn off Google's "social" search results when they enabled that last year (in a previous attempt to revive the moribund Google+). If I'm doing a Google search, it's because I'm asking a question my immediate friends can't help me with.
#DeleteChrome
Is anyone outside of the teenage girl crowd even paying attention to Facebook announcements anymore? I'm legitimately asking. I have a Facebook account that I log into maybe once or twice a year. And most of the circles I spend time in don't really use it much anymore either. Am I the only one that sees Facebook announcements and just shrugs with indifference?
Well, those of us who have successfully managed to stay off facebook can...
It's on America's tortured brow, That Mickey Mouse has grown up a cow
Facebook, search: all the dirty secrets my friends didn't navigate FB's privacy maze properly on.
Or potential employees.
I use Facebook to keep in touch with friends that've moved out of town. Every time they offer a "new exciting privacy feature" it changes my default privacy back to public (from friends only) and I need to redo all my security settings. Now it looks like anything I might miss gets slurped-up in easy to search format.
Fantastic.
What Facebook doesn't seem to realize is that my Facebook "friends" aren't really my friends - they are a large group of family and acquaintances. I don't think my taste in food and/or movies matches maybe 10% of my FB contacts. So if I do search for movies or restaurants my "friends" like, I'm not likely to get any better results than if I search Google.
Plus everyone I know would have to share a lot more information to make this service useful.
I think so. Do you?
I use my profile to bookmark (and share) waht I find interesting. The problem is, if I need to find something over 2 weeks old it takes forever to find it. Why can't I search my own profile?
Slashdot's rate-of-post filter: Preventing you from posting too many great ideas at once.
Zuck said that Facebook spends 3 percent of their CPU power on privacy. With such a low CPU budget dedicated to something as important as users' privacy, it's no wonder they do such a poor job of it.
I read 10%, not 3%:
http://www.engadget.com/2013/01/15/facebook-graph-search/
What would you consider to be a more reasonable amount of CPU budget to spend on excluding search results from some queries? I'm surprised it's as high as 10%, but I never really thought of CPU usage as a metric for privacy protection.
They even got the BBC excited with the news of their news conference. Then the big announcement is this? Really? If it isn't going to turn into big piles of $$ for investors by this afternoon, they better have something earth-shattering coming in real soon. Right now they have a lot of shareholders who are nervous about how much money they lost in a hurry on opening day, and I don't see how this will help them (and I am most certainly glad to not be one of those investors).
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
Rather than blowing it away outright (which some of the comments have done), let's think about it for a sec. There's some cool stuff going on here, and then a big question.
The cool stuff is the technology and innovation. Think about this for a sec - Facebook's engineers are essentially looking at a variety of signals to determine (a) intent and (b) likely outcome. The signals are getting increasingly complex - not simply keyword boolean queries any longer - and, to me, that's a fascinating growth and extension of technology. It's innovation.
It was innovation 10 years ago, now it's what everyone is doing -- Google doesn't do a simple SQL query in a big database to determine the results and ads you see for a query - they mine data from Gmail and their ad network and combine your personal preferences to determine relevance.
Don't kid yourself, 99.9% of these searchers are going to be something along the lines of.. "Girls with mutual friends who became single in the last month" "Single girls near me whose status contains 'drunkkk' more than twice a week" Combine this with imaging searching = awesome "Girls that have dated guys that look like me"
warning: your Facebook privacy settings have been reset
"And what if Facebook “accidentally” resets your settings, like what happened to thousands of users this week?"
It happens, often, to a lot of people. You're either lucky, or lying.
Searching in more advanced forms beyond boolean (semantics etc.) is far far older than facebook - they are doing nothing interesting, indeed it is surprising how little they are extracting with this. "Photos of me when I was 19" - check variable for DOB of user - search photo database tagged with them within a 1 year period. There is very little "innovation" here.
Cpu percentage is hardly meaningful. For example, my site uses 0 cpu on privacy, since it does not collect user data.
"Addressing the obvious privacy concerns, the company said it wouldn't allow users to search content that wasn't already shared with them (or already public). "
Translation:
"This is totally worthless without shared, public data, so we plan to completely fuck with our privacy settings a whole bunch before this rolls out so that we can make sure your data is public and shared."
How can 10% of a server farm go to that? if(notallowed(X,Y)) { etc
How is that notallowed() function written?
boggles my mind. Maybe I am alone and the Ubercoders at FB really can spend 10% of quality CPU time satisfying that func().
H.
Want to know what kind of movies your friends like? ASK THEM.
Want to get set up with a single friend of a friend? ASK YOUR FRIEND.
Want to make Zuckerburg disgustingly rich? USE FACEBOOK.
Only if I can cross referances single girls of friends of friends who post less than or equal to 0 inspirational quotes with pictures per day (including but not limited to: jesus, them being a strong mom, something sappy about relationships, something about being a badass woman, what makes up a real man, their son and/or daughter and how much the love them, etc...)
Facebook: We find out how crazy some of your old friends really are!
How can 10% of a server farm go to that? if(notallowed(X,Y)) { etc
How is that notallowed() function written?
boggles my mind. Maybe I am alone and the Ubercoders at FB really can spend 10% of quality CPU time satisfying that func().
H.
Facebook processes more than 500TB of data a day, and has over 100PB in its Hadoop cluster.
Maybe a simple notallowed() function doesn't scale linearly across many PB of data.
If you've got to perform a search to know which of your friends live in San Francisco, they're not really your "friends".