How can we make sure the RAI can’t tell whether it’s in a simulation?
I’m as sure as I need to be that I’m not in a simulation. (“As sure as I need to be” = I have decided to take no account of the possibility in making any decision.) The argument that any observation I could make could be faked by some all-powerful genie forces more and more improbability onto the hypothetical genie, the more detailed my knowledge of the world becomes, like the person claiming to have a dragon in their garage evading every proposed test. The person making that argument can be replaced by this:
10 INPUT X
20 PRINT "THE GENIE DID IT!"
30 GOTO 10
If I’m smart enough to see that, so is an RAI. That’s not to say that with greater intelligence, it must come to my conclusion. With greater intelligence it is likely to find better arguments to a better conclusion, but if I knew what that conclusion was in advance, it would already be mine.
ETA: A few years back, several car manufacturers made engine control software that could tell when the car was under test for emissions control, and restrict the emissions then, but not on the road. As far as I can gather from Wikipedia, this was deliberately done by the engineers, but I am sure that software programmed to learn how to tune the engine for emissions control under test (“simulation”) and performance on the open road (“reality”) would have learned to do the same thing without any fraudulent intent from the engineers.
How can we make sure the RAI can’t tell whether it’s in a simulation?
I’m as sure as I need to be that I’m not in a simulation. (“As sure as I need to be” = I have decided to take no account of the possibility in making any decision.) The argument that any observation I could make could be faked by some all-powerful genie forces more and more improbability onto the hypothetical genie, the more detailed my knowledge of the world becomes, like the person claiming to have a dragon in their garage evading every proposed test. The person making that argument can be replaced by this:
If I’m smart enough to see that, so is an RAI. That’s not to say that with greater intelligence, it must come to my conclusion. With greater intelligence it is likely to find better arguments to a better conclusion, but if I knew what that conclusion was in advance, it would already be mine.
ETA: A few years back, several car manufacturers made engine control software that could tell when the car was under test for emissions control, and restrict the emissions then, but not on the road. As far as I can gather from Wikipedia, this was deliberately done by the engineers, but I am sure that software programmed to learn how to tune the engine for emissions control under test (“simulation”) and performance on the open road (“reality”) would have learned to do the same thing without any fraudulent intent from the engineers.