3 ms·

One part that's currently missing from Google Glass and that sets it apart from the dashcam example is the ability to continuously record on a circular buffer,

One part that's currently missing from Google Glass and that sets it apart from the dashcam example is the ability to continuously record on a circular buffer, mostly due to constraints in battery life. But once that's possible I would expect many interesting videos to start hitting the internet, as you could essentially save the last x minutes of video to file. A good example of this is the video of the SFO crash landing. It was somewhat unlikely and miraculous that someone just happened to be filming at that time, but in the future there might be tens or hundreds of videos from all the eye witnesses. And not just for this event but any other. It's interesting to think about other repercussions of such technology. From the innocent (such as a large increase in the amount of embarrassing videos on YouTube) to important (such as reconstructing a crime event (think boston bombing)) or scary (blackmail?).

cmbaus 13y ago
This comment just made me realize how right on the original vision of justin.tv was. It was probably 10 years ahead of its time. Maybe PG is a genius at finding founders who see the future.
- icebraining 13y ago
  Or maybe they just looked at the work of Steve Mann, who was doing realtime broadcasting of a wearable camera to the web in '94.
InclinedPlane 13y ago
Technologically it's really only a matter of time before storage gets to a point where keeping months of coninuous 1080p video isn't a big deal, it'll fit on a square cm and cost a pittance. But that's not the scary part. The scary part comes later, due to massive increases in computing power. Imagine a world where, say, computers with terabytes of RAM and thousands of high performance cores are run of the mill server systems. And then it's not just a matter of people storing data but analyzing it, and collating it, deeper than we think is even remotely conceivable today. Face recognition, voice recognition, biometrics, kinematics, writing style, and of course metadata will make public an astounding amount of information. A video of a random group of strangers will be trivial to index and identify every individual through the above techniques. A shockingly detailed account of everyone's lives, where they've been, who they've met, what they've purchased, and so on will be determinable using, eventually, relatively little cost in computing power. Here's an example, let's say there are only a handful of pictures or videos from inside your apartment, imagine if every book on your shelf, every knick knack, every belonging was automatically identified and indexed. It's a world that rapidly spins away from any sort of world we're comfortable or familiar with but we probably won't figure out how to deal with it until well after we've been subjected to the worst abuses and invasions of privacy.
- williamcotton 13y ago
  Time to bust out the scramble suits...
- hrkristian 13y ago
  Then again, computing power and data-requirements both expand at a high rate together, when there are a gazillion cores and terabytes of RAM available for sneaky intelligence business, there will be 4k (or even higher, sky's the limit (and the size of the photon, obviously.) Today various intelligence agencies like the NSA as we now know collect insane amounts of data on the population, but they're unable to effectively sift through it all. When Google Glass gets refined, minimalized and meets fashion, we'll see an impossible amount of increase in data. This is my bright point in the future as far as surveillance goes, for instance dashcams in Russia became a neccessity for the individual citizen to remain safe in face of the law, the results we all know where meteor strikes are documented better than ever, where law officials have to take care what they do, where car accidents are right down to the facts on the mem chip, etc. GGlass can become the same thing on an even more personal level, the truth becomes more easily available when any one person with the ability to press a button can document an event for the world to see.
  - InclinedPlane 13y ago
    It doesn't require processing all data though, just enough of it. Consider something simple like identification. How do you automatically identify someone from video? Face recognition is the go to method now, and it will be used in the future as well, but it will be more capable, of course. But that's only one out of many techniques that are possible, especially with massive amounts of computing power and storage available. Once you've identified someone once in video you can match up other data, such as kinematics and biometrics. How tall you are, how much you weigh, what your overall shape is, your hair color and style, etc. Your typical routines and hangout places. Where you buy coffee, where you watch movies, where you shop. Also, the way you hold yourself, how you walk and stand. Your friends can recognize you without seeing your face, often using these sorts of clues, but there's no reason computers can't use the same techniques. Meanwhile, with enough data computers can catalog your wardrobe. With that sort of information available, and it's really not an enormous amount of data per person, it'll be possible to automatically identify people even without their faces visible, even if they're wearing new clothes and have recently changed their hair styles, and then that info will be cataloged. And again, that's just square one. It'll be almost trivial to keep track of where everyone is, what they are doing, and with whom on a time scale of at least every day. Then there are the scary things. Videos of crowds at protests and rallies can and will be automatically indexed to identify every single participant, even if they're covering their faces. And that's the least imaginative thing that will be possible with such a torrent of data and analysis.

cmbaus 13y ago

This comment just made me realize how right on the original vision of justin.tv was. It was probably 10 years ahead of its time. Maybe PG is a genius at finding founders who see the future.

icebraining 13y ago
Or maybe they just looked at the work of Steve Mann, who was doing realtime broadcasting of a wearable camera to the web in '94.

InclinedPlane 13y ago

Technologically it's really only a matter of time before storage gets to a point where keeping months of coninuous 1080p video isn't a big deal, it'll fit on a square cm and cost a pittance. But that's not the scary part. The scary part comes later, due to massive increases in computing power. Imagine a world where, say, computers with terabytes of RAM and thousands of high performance cores are run of the mill server systems. And then it's not just a matter of people storing data but analyzing it, and collating it, deeper than we think is even remotely conceivable today. Face recognition, voice recognition, biometrics, kinematics, writing style, and of course metadata will make public an astounding amount of information. A video of a random group of strangers will be trivial to index and identify every individual through the above techniques. A shockingly detailed account of everyone's lives, where they've been, who they've met, what they've purchased, and so on will be determinable using, eventually, relatively little cost in computing power. Here's an example, let's say there are only a handful of pictures or videos from inside your apartment, imagine if every book on your shelf, every knick knack, every belonging was automatically identified and indexed. It's a world that rapidly spins away from any sort of world we're comfortable or familiar with but we probably won't figure out how to deal with it until well after we've been subjected to the worst abuses and invasions of privacy.

williamcotton 13y ago
Time to bust out the scramble suits...
hrkristian 13y ago
Then again, computing power and data-requirements both expand at a high rate together, when there are a gazillion cores and terabytes of RAM available for sneaky intelligence business, there will be 4k (or even higher, sky's the limit (and the size of the photon, obviously.) Today various intelligence agencies like the NSA as we now know collect insane amounts of data on the population, but they're unable to effectively sift through it all. When Google Glass gets refined, minimalized and meets fashion, we'll see an impossible amount of increase in data. This is my bright point in the future as far as surveillance goes, for instance dashcams in Russia became a neccessity for the individual citizen to remain safe in face of the law, the results we all know where meteor strikes are documented better than ever, where law officials have to take care what they do, where car accidents are right down to the facts on the mem chip, etc. GGlass can become the same thing on an even more personal level, the truth becomes more easily available when any one person with the ability to press a button can document an event for the world to see.
- InclinedPlane 13y ago
  It doesn't require processing all data though, just enough of it. Consider something simple like identification. How do you automatically identify someone from video? Face recognition is the go to method now, and it will be used in the future as well, but it will be more capable, of course. But that's only one out of many techniques that are possible, especially with massive amounts of computing power and storage available. Once you've identified someone once in video you can match up other data, such as kinematics and biometrics. How tall you are, how much you weigh, what your overall shape is, your hair color and style, etc. Your typical routines and hangout places. Where you buy coffee, where you watch movies, where you shop. Also, the way you hold yourself, how you walk and stand. Your friends can recognize you without seeing your face, often using these sorts of clues, but there's no reason computers can't use the same techniques. Meanwhile, with enough data computers can catalog your wardrobe. With that sort of information available, and it's really not an enormous amount of data per person, it'll be possible to automatically identify people even without their faces visible, even if they're wearing new clothes and have recently changed their hair styles, and then that info will be cataloged. And again, that's just square one. It'll be almost trivial to keep track of where everyone is, what they are doing, and with whom on a time scale of at least every day. Then there are the scary things. Videos of crowds at protests and rallies can and will be automatically indexed to identify every single participant, even if they're covering their faces. And that's the least imaginative thing that will be possible with such a torrent of data and analysis.