I know, the title sounds a bit kooky. But the argument I'm going to put forward is serious, building on work done by David Lewis, perhaps the most influential philosopher of the latter half of the twentieth century (I work as an academic philosopher, so you can trust me when I say that.) I will argue, following Lewis, that the many worlds interpretation of quantum mechanics means that we can be sure that, for each of us, in the future, there'll be someone remembering this moment right now. And that person will have just undergone arbitrarily many torments and tortures stopping just short of death. It's thus reasonable to say that that future person is in hell. But if we all wind up in hell, that seems like a pretty good argument for the existence of the devil.
What I'll do first is introduce very briefly the many worlds interpretation of quantum mechanics in the way it tends to get popularly presented. Then I'll discuss Lewis's paper 'How Many Lives Has Schrödinger's Cat?' and show how it yields the hell conclusion.
THE MANY WORLDS INTERPRETATION (HENCEFORTH 'MWI')
You know Schrödinger's cat. We put a cat in a box rigged up to jug of poison and a hammer. Quantum mechanics tells us that the whether or not a given particle decays in a given period of time is not something we can know for certain; we can only know the probability that it will happen. We put some such particle, which we think has a 50% chance of decaying in the next five minutes, in the box hooked up to the jug and hummer. If it decays, it'll somehow push the hammer, which will smash the jug, leaking the poison and killing the cat. If it doesn't, it won't.
The formalism of quantum mechanics tells us that if there's a 50% chance that a particle is in a certain way W, then there's a quantum mechanical state consisting of it being in W superimposed with it being in ¬W. It's as if the state represents the particle as being both W and NOT-W.
Now how the particle is affects how the cat is: if the particle decays, the cat is dead. So really what we have, looking at the box, is two different superimposed states: W&DEAD CAT and NOT-W&ALIVE CAT.
So the state of the system seems to represent the cat as both alive and dead. Since we never experience such zombie cats, we know something has to give, and on one understanding, what gives us that our observation of the system causes it to resolve into one fixed state.
But there's (at least one) other interpretation of what's going on. If our state represents the cat as both alive and dead, then we should just take that literally. To account for the fact that we never encounter zombie cats, we assume reality splits off any time something like this happens, so that there are a host of worlds each being the outcome of the resolution of a given superposition. (For a much better explanation of this, by someone much better qualified, take a look at https://plato.stanford.edu/entries/qm-manyworlds/.)
HOW MANY LIVES HAS SCHRÖDINGER'S CAT?
These splits occur constantly, not only in special experimental type cases. This lead David Lewis to a horrible thought experiment (http://www.andrewmbailey.com/dkl/How_Many_Lives.pdf; I highly recommend reading this, especially the last few pages.) Consider this variation on Schrödinger's cat. Imagine our cat is sentient in his box. The experiment is slightly different: either the cat is killed, or (rather than going free) it will be non-fatally, but substantially injured. Moreover, the experiment is to be repeated 100 times. What should our sentient cat expect to happen?
It should expect to live through those 100 experiments, gradually getting more and more, but never fatally, injured. Because it eithers lives through them getting injured, or it doesn't. And if it doesn't live through it, then it won't be around to learn the result. The only branches there will be will be non-fatally injured branches. Although it seems to face death 100 times, it can be sure that in 100 experiment's time, he'll be looking back on someone psychologically continuous with him, who has just undergone the experiments. He can be sure, in essence, he's immortal.
This point generalises. Lewis points out (p18) that such consequences aren't limited to fancy poison decay cases. Quantum mechanics is everywhere, and so every chance we face death, there'll be a quantum mechanical probability that we live through it, and so we can guarantee that in the future, we'll look back at ourselves living through it.
So imagine yourself on death's door. Any time now you should expire, you're in pain and ready to leave life. But--you can't! The process of dying will be quantum mechanical, and so there'll be the probability that whatever was about to kill you will in fact not happen, and so you'll find yourself still alive. And this goes on--any time you face death, there'll be a miniscule possibility that you survive.
But that sort of immortality is horrible. As Lewis writes:
"What you should predominantly expect if [the MWI] is true, is cumulative deterioration that stops just short of death. The fate that awaits all of us ...[is to be]...victims of eternal life without eternal health" (How Many Lives Has Schrödinger's Cat?, p20)
CONCLUSION: HELL
That sounds like hell to me. But if you believe the MWI, and many physicists and philosophers of physics do, then hell exists. But if hell exists, that seems like a pretty good reason to think the devil exists. So we should think the devil exists. Which kind of sucks.
P.S.
If you take this seriously, as I am somewhat inclined to, you needn't despair. Indeed, it arguably calls for the opposite of despair: we should hope God exists, in order to prevent this! We should hope God exists to, at one stage, put us out of our misery. While the western theological tradition has been big on God as giver of life, we should take seriously the possibility that he is instead responsible for the sweet release of death the MWI seems to argue is desperately needed.