4 ms·

So like, how exactly did this accidentally happen? It's not like skynet's running those vans and just decided to sniff the traffic of open wifi nets it passed.

• X-Istence 16y ago
  The code was compiled with -DDEBUG and instead of recompiling without it was put on the vans driving around. Honest mistake, they are owning up to it, and removing the data. The amount of data that they could have picked up while moving around driving is very little and it is available to anyone with a wifi antenna and some time to drive around. I'd be less worried about Google than about the kid a couple doors down that just found out about nmap, wireshark and kismet.
  • keefe 16y ago
    Either I skimmed that article very poorly or you have some inside information on this... I may be slightly paranoid but I am always skeptical about stuff like this. I could imagine a lot of useful pieces of information that could be gleaned, like a geocoded map of where people are ACTUALLY using google and for what... I don't really know, it just gives me a little bit of a funny feeling
    • X-Istence 16y ago
      The guy asked how exactly did this happen, I was speculating. I am in no way associated with google, and a quick Google of my username would have figured that. From the quick skimming of the articles I did I saw several mentions on different places of debug versions. It is very plausible. You may be paranoid, but if you had read any of the articles you would have seen that the cars channel hop 5 channels a second, and that they are constantly driving. All they are looking for is beacon information sent out by access points. Apparently this also captured a couple of packets of actual data that was not discarded by the computers in the Google cars because of a debug build (from what I have read, again that might be pure speculation, but makes sense). That won't be enough time to even figure out if the user is using Google or not. Go grab an wifi antenna, a wireless networking card, your favourite Linux distribution and go wardriving. You will pick up the same information that Google picked up. Now, if from the small samples you did get you can determine where people are using Google and for what and can make it valid research, good on you, but I doubt you will get much more than the occasional URL, and if you are lucky that URL is a Google URL.
      • keefe 16y ago
        You're assuming that this business about debug modes and the cars channel hopping so quickly is not a cover story. There's no way for us to know that. My theory was that given a bunch of interesting data and no way to see what they were really doing, there's no telling what was really going on.
        • X-Istence 16y ago
          Sure, but if all of us are walking around with tinfoil hats on, looking over our backs for the men in black suits, all of the time we wouldn't get much done now would we? Also, once again, the channel hopping is fairly normal for a wifi wardriving setup, feel free to go test this theory on your own, as well as go driving down the road at 25 - 50 Mph and get anything but a few packets of data from the wireless networks you do come across where the owners at the moment you are driving by are using their wireless, and data is being transmitted in clear text.
          • keefe 16y ago
            and if all of us accepted the stories of the corporate and government Powers then they could do whatever they like. There's an awfully big gulf between being skeptical of official stories and being so paranoid that it interferes with productivity. I'm certainly not interested enough to test any of this stuff... but my thought was OK, could they intercept GET requests to URLs off of google.com and that gives query strings etc? Even 100 URL+GEOCODE per hour would be win. I'd certainly be tempted to try it if they were my vans, if for no other reason than to see if it was possible. I'm doubting that the range is 25-50mph in residential zones btw.
• rmorrison 16y ago
  Yeah, it's time to start using duckduckgo.com exclusively = )