Robert Scoble discusses Google Glass patents for gaze-tracking and contextual APIs that detect what you're doing: walking, running, driving, shopping. The system could infer intent from where you look in a store window and your economic status, then serve targeted ads and product suggestions.
All sorts of fun stuff. But for example, just a couple of days ago, one of the things that really freaked out some people online was a patent filed by Google about sort of watching your gaze and estimating which arcs are which ads work and which ads don't work and where they should be placed so that you get to see them more and more. So how about that business model with ads?
I think it goes a lot deeper than that. I think Google is building a contextual os, an operating system that knows your context. And what do I mean by that? Already they have an API that knows whether you're walking, running, biking or driving. Okay, that's four contexts. Next year it'll be 8 and the year after it'll be 16 and the year after it will be 32. So think about, can you figure out somebody shopping just by watching them walk around? Of course you can. Because if you're in a, first of all, you're in a shopping mall. That means you're shopping across the geofence into the mall. Second of all, this thing knows where you're aiming. So if you're running through the mall, first of all, it knows that you're running. And so it can say, okay, he's running, he's not shopping. But shoppers walk through the mall and all of a sudden they turn and look at a store, their intent changed, their gaze changed. Window shopping, Window shopping, right? If you're at Macy's in New York, you're looking at the store, at the windows, you're not walking down the street going past it, right? And so it can know right there, it can switch, oh, you're in a shopping context. And then it can say, then it can fire the camera, take a picture of that store window and then watch where you look. And if it knows anything about that window, if it knows there's a trumpet in the upper right and there's a drum in the upper left and there's a guitar in the middle and there's a saxophone in the middle left. If you look at the trumpet, it knows. It can put that in the database. Oh, this guy's thinking about a trumpet for some reason. And then it would match up with other contexts. Do you have a kid? Oh, you might. Oh, is there a birthday next week? It could start making some assumptions about why you're shopping and it can even ask you.
But those assumptions are a double edged sword, aren't they?
Yeah, they are. You could have false positives, but you could ask the user, you could say, hey, we noticed you were interested in that trumpet, are you? Would you like to see other models? Would you like to know other stores nearby? Would you? Because it can try to guess what your intent is, right? And certainly Google is watching everything, right? So if you go to Google and you type trumpet, all of a sudden it knows you're really strong attached to trumpet. And if you start looking for music stores or if you ask Google Glass, ok, glass, where's the closest music store, all of a sudden it knows your intent is really strong on music and trumpets, right? And then you're going to start showing it some brand names and watch your behavior and it'll know, by the way, it'll know your economic circumstances, right? So if you're a billionaire, it's going to show you a different kind of trumpet than if you're Robert Scovel and my friends who are poor or who aren't as well off, it'll show them a far different trumpet. So it'll put you in a different category. If you're part of a band, it'll know a lot more. Or it might look at if it knows what you're buying it for, then if it's for yourself, it'll watch your Spotify behavior, right? So it'll know, oh, you're into this kind of music, you're into jazz music. You should look at these trumpets because these are more appropriate for jazz than classical.
