On Sun, Jan 19, 2014 at 4:17 PM, Russell Standish <[email protected]>wrote:
> If a rational agent can compute its utility to determine its next course > of action, then so can any observer with access to the same > environmental information. > Yes, but only by going through the same process the rational agent does. If the observer is using hardware that runs at the same speed then the observer will finish his calculation and know what the rational agent is going to do at the same instant the rational agent actually does it. If the observer had not done a single calculation and had just stuck to observing he would have found out what the rational agent is going to do just as quickly; all the calculations he made were pointless. For a prediction to be a prediction it must tell you what's going to happen BEFORE it happens, if you can't do that with a rational agent then that agent is unpredictable. Of course if the observer DID have faster hardware then he could make a successful prediction, provided he didn't tell the rational agent what that prediction was; if he did spill the beans then you'd need to simulate both the rational agent and the observer on hardware that was faster still to figure out what the rational agent would do before he actually did it. > Its got nothing to do with the Halting problem. > It has lots and lots to do with the Halting problem. John K Clark -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/everything-list. For more options, visit https://groups.google.com/groups/opt_out.
