I will execute this algorithm unless doing so produces non-halting states, in which case I will only take Box B since this creates a schelling point at my next most preferred worldstate.
At least two issues there:
halting problem is famously undecidable—in general, there is no way to know whether you are in an infinite loop vs a very long finite loop (but you can approximate with a timeout)
you presuppose some way to find the best fallback decision—how does that work? Almost like you are presupposing you already know how to slove the problem?
At least two issues there:
halting problem is famously undecidable—in general, there is no way to know whether you are in an infinite loop vs a very long finite loop (but you can approximate with a timeout)
you presuppose some way to find the best fallback decision—how does that work? Almost like you are presupposing you already know how to slove the problem?
My understanding is that you don’t solve the question “will it halt?” but will it halt in the time I’m given in the problem at hand.
That’s exactly what I meant by “you can approximate with a timeout”.