I’ve been thinking about ideas along these lines a bit recently. I think, if this kind of thing is possible (i.e. branches of low probability never go to zero, and are always realizable), it would be important to make sure that all of your probability mass is crisply divided with some going to universes where all consciousness is quickly extinguished and some to universes that continue with whatever properties you want, with little to none leftover for other possibilities.
In other words, quantum suicide is a bad idea for the individual (you leave around mourning loved ones), but perfectly fine and acceptable to exploit on the civilization level.
In other words, quantum suicide is a bad idea for the individual (you leave around mourning loved ones), but perfectly fine and acceptable to exploit on the civilization level.
No, it is not. Quantum probabilities enter expected utility calculations. If civilization is destroyed with high quantum probability, expected utility plummets.
Quantum probabilities enter expected utility calculations
Well of course they do, but the question is whether it should be on a relative or absolute basis. That is, do you care about the probability along all futures, or just the relative probabilities among ones where there will be conscious observers? My suggestion above is from the relative position, while it sounds like your preference is to calculate utilities according to the absolute position.
I do not think it is a settled question which is the objectively right way to do it. I lean towards relative currently, but even stating that doesn’t narrow down exactly how to calculate expected value (is there weighting by number of observers? kind of observers?). I think we are all still confused on all of this, as evidenced by all the discussion here on the topic of anthropic probabilities, magical reality fluid, L-zombies, etc.
I’ve been thinking about ideas along these lines a bit recently. I think, if this kind of thing is possible (i.e. branches of low probability never go to zero, and are always realizable), it would be important to make sure that all of your probability mass is crisply divided with some going to universes where all consciousness is quickly extinguished and some to universes that continue with whatever properties you want, with little to none leftover for other possibilities.
In other words, quantum suicide is a bad idea for the individual (you leave around mourning loved ones), but perfectly fine and acceptable to exploit on the civilization level.
No, it is not. Quantum probabilities enter expected utility calculations. If civilization is destroyed with high quantum probability, expected utility plummets.
Well of course they do, but the question is whether it should be on a relative or absolute basis. That is, do you care about the probability along all futures, or just the relative probabilities among ones where there will be conscious observers? My suggestion above is from the relative position, while it sounds like your preference is to calculate utilities according to the absolute position.
I do not think it is a settled question which is the objectively right way to do it. I lean towards relative currently, but even stating that doesn’t narrow down exactly how to calculate expected value (is there weighting by number of observers? kind of observers?). I think we are all still confused on all of this, as evidenced by all the discussion here on the topic of anthropic probabilities, magical reality fluid, L-zombies, etc.
I’m sure you’ve read it, but to give everyone else context, here’s the paper on civilization-level quantum suicide.
Also very relevant to this discussion is the paper “Many-Worlds Interpretations Can Not Imply ‘Quantum Immortality’” [pdf]
Edit: Original link wasn’t working, so I uploaded a copy.
I haven’t read it, thanks! Unfortunately, that link doesn’t seem to go to the intended destination.