Why is the discrimination problem “unfair”? It seems like in any situation where decision theories are actually put into practice, that type of reasoning is likely to be popular. In fact I thought the whole point of advanced decision theories was to deal with that sort of self-referencing reasoning. Am I misunderstanding something?
If you are a TDT agent, you don’t know whether you’re the simulation or the “outside decision”, since they’re effectively the same. Or rather, the simulation will have made the same choice that you will make.
If you’re not a TDT agent, you gain more information: You’re not a TDT agent, and the problem states TDT was simulated.
So the discrimination problem functionally resolves to:
If you are a TDT agent, have some dirt. End of story. If you are not a TDT agent, I have done some mumbo-jumbo, and now you can either take one box for $1000 or $1m, or both of them for $1001000. Have fun! (the mumbo-jumbo has nothing to do with you anyway!)
Why is the discrimination problem “unfair”? It seems like in any situation where decision theories are actually put into practice, that type of reasoning is likely to be popular. In fact I thought the whole point of advanced decision theories was to deal with that sort of self-referencing reasoning. Am I misunderstanding something?
If you are a TDT agent, you don’t know whether you’re the simulation or the “outside decision”, since they’re effectively the same. Or rather, the simulation will have made the same choice that you will make.
If you’re not a TDT agent, you gain more information: You’re not a TDT agent, and the problem states TDT was simulated.
So the discrimination problem functionally resolves to:
If you are a TDT agent, have some dirt. End of story.
If you are not a TDT agent, I have done some mumbo-jumbo, and now you can either take one box for $1000 or $1m, or both of them for $1001000. Have fun! (the mumbo-jumbo has nothing to do with you anyway!)