> Google prototypes have a big, red stop button that, > I assume, takes precedence over hackable software. You think so? I would assume not. I doubt the ca

> Google prototypes have a big, red stop button that, > I assume, takes precedence over hackable software. You think so? I would assume not. I doubt the car is just going to screech to a halt via some hardware switch, think about the danger of someone pressing that while in the center lane of a highway. Instead it's probably synonymous with giving the software a "stop now" command. I.e. it would know to safely pull over to the side of the road, or take the next off-ramp in that situation. So if someone had hacked the driving software that could be overridden. I do think idibidiart is being overly alarmist. Self-driving cars don't have to be perfect, and there could certainly be malicious hacking incidents, but they're highly likely to outnumber the cases where people die now due to imperfect human drivers.