4 ms·

You know what's most surprising about this? That the developers, knowing fine well it was for a single target and the damage it could do in the wild, didn't imp

You know what's most surprising about this? That the developers, knowing fine well it was for a single target and the damage it could do in the wild, didn't implement a kill switch. Quite frankly cowboy coding like that is why we'll end up with Skynet becoming self-aware.

46Bit 14y ago
How do you put a kill switch in? The entire point was for it to be on an offline system, and the traffic for it checking for Internet to in turn check for an order like that might have given the game away. It sounds like they had something to check it wasn't outside and someone just screwed up.
- nicholassmith 14y ago
  You engineer a solution. Quite frankly if you can't think of how to put a kill switch into something like this then it shouldn't be getting written or allowed into the wild. You can't add that to Bugzilla and hope to fix it later.
- wpietri 14y ago
  I think the obvious thing would be to look for Internet activity on the machine. E.g., recent files for popular sites in a browser cache directory. Or live network traffic for public netblocks.
  - Symmetry 14y ago
    But it has to start out on machines connected to the public internet before spreading to the secret facility via USB drive. If it didn't spread when connected to the Internet, it wouldn't work. And it already doesn't blow up centrifuges when it's not on the target network.
    - MartinCron 14y ago
      it already doesn't blow up centrifuges when it's not on the target network I can verify that. None of my centrifuges have been blown up.
talmand 14y ago
Actually, in this case, a kill switch would be a bad idea for the original coders. It opens the possibility of the target rendering the attack pointless. Imagine a tank with a nice big, red button that shuts down the internal systems instantly. Now imagine that button on the outside of the tank. Sure, you could put some kind of password encoded lock on the button but it's a huge START HERE sign on the outside of the protection of the tank.

46Bit 14y ago

How do you put a kill switch in? The entire point was for it to be on an offline system, and the traffic for it checking for Internet to in turn check for an order like that might have given the game away. It sounds like they had something to check it wasn't outside and someone just screwed up.

nicholassmith 14y ago
You engineer a solution. Quite frankly if you can't think of how to put a kill switch into something like this then it shouldn't be getting written or allowed into the wild. You can't add that to Bugzilla and hope to fix it later.
wpietri 14y ago
I think the obvious thing would be to look for Internet activity on the machine. E.g., recent files for popular sites in a browser cache directory. Or live network traffic for public netblocks.
- Symmetry 14y ago
  But it has to start out on machines connected to the public internet before spreading to the secret facility via USB drive. If it didn't spread when connected to the Internet, it wouldn't work. And it already doesn't blow up centrifuges when it's not on the target network.
  - MartinCron 14y ago
    it already doesn't blow up centrifuges when it's not on the target network I can verify that. None of my centrifuges have been blown up.

talmand 14y ago

Actually, in this case, a kill switch would be a bad idea for the original coders. It opens the possibility of the target rendering the attack pointless. Imagine a tank with a nice big, red button that shuts down the internal systems instantly. Now imagine that button on the outside of the tank. Sure, you could put some kind of password encoded lock on the button but it's a huge START HERE sign on the outside of the protection of the tank.