Károly Zsolnai-Fehér explains Nick Bostrom's argument that one of three things must be true: advanced civilizations die out before reaching maturity, they choose not to run ancestor simulations, or we're probably living in one. Bostrom's logic: if mature civilizations can simulate reality, they'd run many of them, making it statistically likely we're inside one.
Dear fellow scholars, this is two minute papers with and it is time for some minds to be blown. We're going to talk about a philosophy paper before we start. A quick an ancestor simulation is a hypothetical computer simulation that is detailed enough that the entities living within are conscious. Imagine a computer game that you play that doesn't contain mere digital characters, but fully conscious beings with feelings, aspirations and memories. There are many interesting debates among philosophers on crazy, elusive topics like prove to me that I'm not in a dream or I'm not just a brain in a bottle somewhere that is being fed sensory inputs. Well, good luck. In his paper, Nick Bostrom, philosopher, offers us a refreshing take on this topic and argues that at least one of these three propositions is is true. Almost all advanced civilizations go extinct before achieving technological maturity. There's a strong convergence among technologically mature civilizations in that none of them are interested in creating ancestor simulations. And here's the we are living in a simulation. At least one of these propositions is true. So if you say no to the first two, then the third is automatically true. You cannot categorically reject all three of these, because if two are false, then the third follows. Also, the theory doesn't tell which of the three is true. Let's talk briefly about the first one. The argument is not that we go extinct before being technologically advanced enough to create such simulations. It means that all civilizations do. This is a very sad case, and even though there is research on the fact that war is receding, there's a clear trend that we have less warfare than we've had hundreds of years ago. I've linked a video on this here from Kurzgesagt. It is still possible that humanity eradicates itself before reaching technological maturity. We have an even more powerful argument that maybe all civilizations do such a crazy proposition. Second point. All technologically mature civilizations categorically reject ancestor simulations. Maybe they have laws against it because it's too cruel and unethical to play with sentient beings. But the fact that there is not one person in any civilization, in any age who creates such a simulation. Not one criminal mastermind anywhere, ever. This also sounds pretty crazy. And if none of these are true, then there is at least one civilization that can run a stupendously large number of ancestor simulations. The future nerd guy just goes home, grabs a beer, starts his computer in the basement and fires up. Not a simple computer game, but a complete universe. If so, then there are many more simulated universes than real ones. And then with a really large probability we're one of the simulated ones. Richard Dawkins says that if this is the case, we have a really disciplined nerd guy because the laws of physics are not changing at a whim, we have no experience of everyone suddenly being able to fly. And as the closing words of the paper, with graceful eloquence in the dark forest of our current ignorance, it seems sensible to apportion one's credence roughly evenly between 1, 2, and 3. Please note that this discussion is a slightly simplified version of the manuscript, so it's definitely worth reading the paper. If you're interested, give it a go. As always, I've put a link in the description box. There's no conclusion here. No one really knows what the answer is. This is open to debate and this is what makes it super interesting. And now my personal opinion, it's just an opinion. It may not be true, it may not make sense, and may not even matter. Just my opinion. I'd go with the second. The reason for that is that we already have artificial neural networks that outperform humans on some tasks. They are still not general enough, which means that they are good at doing something like the Deep Blue is good at chess, but it's not really useful for anything else. However, the algorithms are getting more and more general, and the number of neurons that are being simulated on a graphical card in your computer are doubling every few years. They will soon be able to simulate so many more connections than we have, and I feel that creating an artificial super intelligent being should be possible in the future that is so potent that it makes a universe simulation pale in comparison. What such a thing could be capable of. Hmm, it's already getting too long. I just can't help myself. You know what? Let's discuss it in a future two Minute Papers episode. I'd love to hear what you fellow scholars think about these things. If you feel like it. Please leave your thoughts in the comment section below. I'd love to read it. Thanks for watching and for your generous support and I'll see you next time.
