Facebook called everyone to Menlo Park this morning for one of those invite-only press things, and by lunchtime my Twitter feed was just wall-to-wall screenshots of the new search bar. They're calling it Graph Search, and Zuckerberg apparently stood up there and called it the third pillar of the site, right up there with News Feed and Timeline. Big words for something that's not even out yet — it's a waitlist signup today, not a real launch.
I've read through maybe a dozen writeups over lunch and the demo queries are the part that stuck with me. Things like "restaurants my friends have been to in Delhi that aren't sushi places" or "friends of friends who like hiking and live in Seattle." It's clever, I'll give them that. It's not web search, it's not trying to compete with Google for "best pizza near me" — it's entirely built out of stuff you've already told Facebook, structured into a giant graph of people, places, photos and likes, and now you can actually query it in plain English instead of just scrolling.
And that's exactly what makes me uneasy about it, if I'm honest.
I spent about twenty minutes after reading the announcement going back through my own profile checking what I'd tagged and liked over the last two years, because up until today none of that stuff was really searchable in any meaningful way. Sure it existed. Sure a determined ex or a nosy coworker could theoretically scroll back far enough. But scrolling is friction, and friction matters more than people give it credit for. A status I liked in 2011 because a friend posted something funny was basically dead the moment it scrolled off my feed. Once Graph Search rolls out for real, "things Ramith has liked, posted by people who aren't his friends" becomes a query some rando could actually type in, assuming the privacy settings actually hold (they say they'll respect existing privacy settings, for what that's worth).
I'm not trying to be alarmist about this, and I bet in six months we'll all be using it constantly and forgetting why it seemed weird. That's basically what happened with Timeline last year: everyone panicked for two weeks, dug through old photos, untagged a bunch of stuff, and then never thought about it again. I'll probably do the exact same dance with Graph Search once it actually opens up past the waitlist.
But there's a difference between Timeline reorganizing stuff that was already visible to whoever could already see it, and Graph Search making that same stuff findable on purpose, by strangers, on demand. Those aren't the same kind of exposure even if Facebook wants to lump them together as "it's all information you already agreed to share."
The other thing nobody in the coverage I've read is talking about much: this whole thing depends on Facebook's own like and tag data being decent, and anyone who's used the site more than a year knows how messy that data actually is. Half my "likes" are pages I clicked once back in 2011 for a contest entry that I don't even remember entering. I'm not sure "people who like the same forty random pages as Ramith" produces anything useful, but someone will run that query eventually, probably a recruiter or an ex, and I genuinely have no idea what it turns up.
Small side complaint while I'm here: every single article covering this today used a screenshot of the exact same demo slide, the one with the hiking friends in Seattle. Not one outlet bothered to try their own example query before publishing. Lazy tech journalism isn't new, but it's a little funny watching it happen in real time across a dozen tabs at once.
Anyway. I've requested access to the waitlist because of course I have, I'm not made of stone, and I'll write something up properly once it's actually usable instead of just guessing off a keynote. For now I'm doing what apparently a lot of people are doing today, which is quietly going back through two years of tags I'd forgotten existed. If you're on Facebook at all, it might be worth twenty minutes of your afternoon too.