Aug 8, 2018
There is a saying, especially when quantum mechanics is concerned, that everything adds up to normality. Here I argue that is not entirely correct.
The word "ontology" in the title refers to our conception of the basic building block of reality. In quantum mechanics the ontology is the wave function, in general relativity it is spacetime.
This idea assumes the many-worlds interpretation of quantum mechanics is the correct one. In this interpretation there is an infinity of universes which begin as the same. When an event happens it can go one way in some universes (Schrödinger's cat dead) and another way in another (Schrödinger's cat alive). The number of universes stays constant, they just get diversified.
The laws of general relativity allow time travel. Time travel is considered impossible since it leads to certain paradoxes, such as the grandfather paradox and is hence deemed logically inconsistent. Those paradoxes are resolved if the time traveler travels not only in time, but also to another universe, as David Deutsch explains in his book The Fabric of Reality and elsewhere. I'm not a physicist but this would imply that the laws of quantum mechanics allow for travel between universes.
The fundamental laws of physics are not known but whatever they are the Earth will still be round and single photons will still interfere in the double-slit experiment. Whatever the true laws of physics are, they will probably include parallel universes. It is not so clear if they will allow for travel between those universes and if they allow, will such travel be practically feasible.
There are three basic possibilities with regards to the ultimate laws of physics and inter-universe travel:
Since we don't know what the ultimate laws of physics are, we can only assign probabilities to each scenario. A utilitarian who assumes a non-trivial probability to the possibility of inter-universe travel should think about the consequences of using quantum random number generators (QMRNG). The use of QMRNG causes quantum diversification.
Let's say we have 10 universes which are all identical, they all have you in them, you are tied to the tracks and a trolley is approaching. You have two buttons to press. Button A in all universes has the same effect but you are not sure which the effect is, there is a 90% of it not doing anything and 10% of it stopping the trolley. Button B uses a QMRNG and stops the trolley in 1 universe while letting it run you over in 9 universes. To a utilitarian the total expected utility from pressing any button is the same. In case A, the expected utility for each universe is 0.1 lives saved, so for total we get 10 * 0.1 = 1 life saved. In case B, the total expected utility is 1 life saved. The expected utility is the same, except... if inter-universe travel is possible and you are an expert surgeon which can save your copy's life after it has been run over. In that case you survive in one universe and travel to other universes one by one and save the other copies. Taking the sum of utility of all universes for all times, the situation when a QMRNG is used looks a lot different than when not used. When not used, at one point in the future, the utility becomes zero and stays zero. When used, you can recover. This applies to existential risk if we substitute "our copy" with "our entire species" and "revival" with "repopulation".
Let's say that in the moment you are pushing the button you don't know if inter-universe travel is possible. There is a non-zero probability p of it being possible. As long as it does not cost you anything, the expected utility of pressing the button B (which we can write as EU(B)) is always higher, it is (1 - p) * EU(A) + p * (EU(all alive)). If there is a cost C, we just subtract it from the result. In case C is too high and p is too low, it can be better to press the button A.
We can assume that inter-universe travel consumes some resources and takes some time, perhaps there is also a constraint that universes you travel to need to be similar enough to your own. This would limit the speed of our travel through the configuration space (the linear space in which each universe is a point) and also limit the range of such travel. The universes naturally have a certain degree of diversification - events in some universes go one way, in others go another way. It is not clear (well, at least to me) what is the extent of this diversification. When walking through the city I may be unsure should I go left or right. It could be the case that I go left in 1% of the universes and right in 99% of them, or I go left in 5%, or any other percent, the closer the percent being to 50% the higher the diversification. It could also be the case that in 100% of situations I go right, and only rarely is there a decision I make differently in different universe, with most my decisions being the same in all universes. As we know from chaos theory there are systems which are highly sensitive to initial conditions, such as the weather, and they could introduce diversification - initially the changes between some universes are small but they get amplified with time.
Let's say there is an astronomically high proportion of universes in which homo sapiens went extinct and only a small proportion in which it didn't. It would be better to increase the proportion of survivors, since the travel to extinct universes in that case would be faster. Also, increasing the number of survivor universes means that survivors will be spread out in configuration space and as such they will be able to repopulate a larger area of configuration space. This implies that increasing the level of quantum diversification through the usage of quantum mechanical random number generators reduces existential risk.
The effect of using QMRNG could still be negligible if: