A good and logical counter-argument. I respond with this. What about a non-google solution?
http://www.youtube.com/watch?v=Flvd5gVT7fg ...but seriously, a Google Glass-like apparatus independent of Google is technology of an entirely different order - only a handful of companies can meaningfully augment reality with internet data. Now if you're talking about something to the effect of "wearable video camera but stores everything to a MicroSD card", then most of the other issues still apply - no notice, and questionable motivation by the wearer.
Something you have much more total control of the data that flows in and out of it.
The use of the second person here is the crux of the issue: is "you" referring to "the wearer", or "the individuals being recorded"? If the former, the above issues are still in force, motives are still questionable, and it's entirely possible that one of the motives is "uploading to Google by hand". If "the individuals being recorded", then the implementation gets messy, and even if that is somehow solved, then if you have two people recorded, one wants the video uploaded to Facebook and the other does not, who wins?
Is it ok if i do face-matching on my own hardware to build a better picture of who i come near day-to-day?
Same problems - questionable motives of the wearer. Now if I were somehow confident that your only intention was to see how many people you pass by on the street more than once in a week, then yeah, I wouldn't have a problem with it. If you're a marketing rep for Clearchannel counting that data in order to directly state the number of impressions a given video billboard will get, then I'm a bit less enthused. This completely sidesteps the "think of the children" and "rule 34" issues presented...and even if I asked a given Glass wearer what their motive was, and assumed they were 100% truthful, I won't know who's the curious one who doesn't trust anything to the cloud, and who's the sketchy person who streams to Google by choice, until it's far too late.
Again, this is where messes come in. "with who", "why", "what will happen to it when it's over", and "even if I trust you, the other five people I do not" issues are still at play.
Amongst the issues is how we define "aggregate". If by "aggregate" you mean "compare my own footage with...my own footage", then sure...back in my day, we called that 'video editing'. Tying that in with a database of Facebook profiles? uploading data for distributed analysis amongst 10,000 other glass wearers who are curious as to who passes the same people in a day? Selling videos of city life as royalty free B-roll for news broadcasters?
If you're blurring everyone's face for the sake of anonymity, then it kinda defeats the purpose of videotaping in the first place in most respects.
Herein lies the true issue at hand. None presently exist. There is no societal construct, no precedence, and no feasibly-enforceable-legislation to adequately define the lines. Thus, we end up with what loosely amounts to anarchy - "I", the person being recorded, have a certain set of standards, and "you", the person recording, have another. When these collide, who wins, and why? In a 1:1 situation like that, I guess the answer would be "the one who's most imposing", which is a pretty poor set of circumstances under which to define what's socially acceptable. I'm certainly not opposed to the questions being posed, and discussions being had, but like I said - this is the heart of the matter.