5 ms·

What bothers me about the idea of an ancestor simulation is the suggestion that a computer which exists inside the universe could simulate the universe with eno

What bothers me about the idea of an ancestor simulation is the suggestion that a computer which exists inside the universe could simulate the universe with enough fidelity that it could actually serve as a useful ancestor simulation. An ancestor simulation seems to be a contradiction, like a box that contains itself within.

sudosysgen 7y ago
It's still possible, but you will have statistical inaccuracies. If your analysis shows that say almost all events in this area can be abstracted up to this value without losing a significant amount of accuracy that makes the inner box smaller. Factor in our limited speed of travel and the one way nature of causality relative to us on the cosmic scale, you get a theoretically feasible simulation.
- simonh 7y ago
  It _may_ be possible, and the inaccuracies in such a simulation _might_ be useful to meet some goal, if a valuable use could be determined that would justify the costs involved, but there's nothing inevitably true about any of those propositions. It's a bit like the Drake equation, reasonable people can assign different probability values to the various factors and come to radically different conclusions. For example if I think there's a very low chance a valuable enough use for such a simulation could be found, and there's a low chance such a simulation would meet the criteria to be useful, and that there's a low chance such a simulation would need to be so accurate as to have actually conscious simulated inhabitants to meet those goals, then the chance that we are in a simulation comes to be vanishingly small.
afpx 7y ago
I didn’t read the paper, only the blog post. But, why does the simulation have to exist within the same universe? Also, wouldn’t a procedurally-generated universe alleviate the need for high-fidelity?
- root_axis 7y ago
  > why does the simulation have to exist within the same universe Where else could it exist? > wouldn’t a procedurally-generated universe alleviate the need for high-fidelity? Could you elaborate on this, I'm not exactly sure what you mean.
downerending 7y ago
One way out is to realize that such a simulation doesn't really need that much fidelity. It just needs enough to seem like it does. We've all had the experience of dreams that seemed very "realistic", only to awake and realize that they didn't really make that much sense.

sudosysgen 7y ago

It's still possible, but you will have statistical inaccuracies. If your analysis shows that say almost all events in this area can be abstracted up to this value without losing a significant amount of accuracy that makes the inner box smaller. Factor in our limited speed of travel and the one way nature of causality relative to us on the cosmic scale, you get a theoretically feasible simulation.

simonh 7y ago
It _may_ be possible, and the inaccuracies in such a simulation _might_ be useful to meet some goal, if a valuable use could be determined that would justify the costs involved, but there's nothing inevitably true about any of those propositions. It's a bit like the Drake equation, reasonable people can assign different probability values to the various factors and come to radically different conclusions. For example if I think there's a very low chance a valuable enough use for such a simulation could be found, and there's a low chance such a simulation would meet the criteria to be useful, and that there's a low chance such a simulation would need to be so accurate as to have actually conscious simulated inhabitants to meet those goals, then the chance that we are in a simulation comes to be vanishingly small.

afpx 7y ago

I didn’t read the paper, only the blog post. But, why does the simulation have to exist within the same universe? Also, wouldn’t a procedurally-generated universe alleviate the need for high-fidelity?

root_axis 7y ago
> why does the simulation have to exist within the same universe Where else could it exist? > wouldn’t a procedurally-generated universe alleviate the need for high-fidelity? Could you elaborate on this, I'm not exactly sure what you mean.

downerending 7y ago

One way out is to realize that such a simulation doesn't really need that much fidelity. It just needs enough to seem like it does. We've all had the experience of dreams that seemed very "realistic", only to awake and realize that they didn't really make that much sense.