All right, Sam is away today, and you know, when the cat's away, the mice will play. So I get to indulge in some of my favorite rabbit holes I go down.
Our guest today is both an entrepreneur. You've built and sold multiple companies in the video gaming space, which kind of led you down the pathway of trying to think through: are we living in a simulation? So let's start there. What is the theory, and then how you kind of stumbled into believing this?
Yeah. So simulation theory is the idea that the world that we're living in, the physical world, is not really physical at all. That in fact we are inside of some type of virtual reality or a computer program. I like to say we're inside a massively multiplayer video game. The most popular pop culture reference really is The Matrix. Neo starts off at the beginning going to work at a computer software company called MetaCortex, and he thinks it's a real world and he thinks he has a real job, and it turns out the whole thing was an elaborate, you know, virtual reality of the year 1999, which is of course when The Matrix came out. So that's kind of a very quick overview.
Yeah. And I know it sounds, you know, obviously a little bit off-the-wall. I remember the first time I heard about it. I think Elon Musk was talking about it like a decade ago, and he says it in the very Elon matter-of-fact way where he's like, you know, if you look at the plot of video games, we go from Pong, which is like two lines and a dot on a screen, to where we have, you know, video games today. That's like a 30, 40-year period, which is like the blink of an eye in the time scale of Earth. So 30, 40 years, we go from two lines and a dot to like, you know, Grand Theft Auto type of stuff. And then he said, you know, obviously any sufficiently technologically advanced civilization would want to create a simulation because it's way cheaper to find out how things are going to go in a simulation than it is to go test things in the real world. So there's huge incentive. Clearly, if you could, you would build it. And so then there's a question of like, is this the real one and we're going to make a simulation later, or did some, you know, was there a real one and this is the simulation? And his conclusion was more than likely this is the simulation. And he just sort of said it matter-of-fact, and then, you know, they changed topics, and my mind was blown, and, you know, my brain is scattered on the wall, and I'm trying to figure out what the hell he's talking about. I started thinking about it, and, you know, at the time I would say this was pre-AI when he was talking about this. It sounded like, you know, kind of just nerd dream. I don't know, it didn't, you know, it was hard to really wrap your head around it. There wasn't a lot where you could sort of like squinch your eyes and see it. Then now let's fast forward. We have AI chatbots, AI voice, AI imagery. I mean, you can't tell if a video is real or fake anymore. Like we're already at that cusp, the Turing test of looking at a video. My mom can't tell if a video is real or fake at all. Zero.
Yeah. I mean, is my background real or not at this point, right? Similarly, am I really talking to you or not? Actually, I'm not. I'm talking to my computer, which is transmitting the image, and it's, you know, transmitting it to your computer and playing it out on your computer. But Elon said that back in the year 2016 at the Code Conference, and his conclusion was that the chances that we are in base reality, which would mean you're not in a simulation, is one in billions. So you flip that around and say, okay, what he was trying to say is the chances that we are in a simulation is billions to one, or...
So what is the history of this? Like, where did he get it? Was this a popular theory I just hadn't heard about? You know, what's the kind of origin story?
Well, the origin story, what he was referring to is something called the simulation argument. Now, what happened was The Matrix came out in 1999. This wasn't the first, you know, representation of living in a virtual world, but it became the most talked-about film of that year. Before that, you had Philip K. Dick, who's a well-known science fiction writer behind, he wrote the novels behind Blade Runner, Minority Report, and the series The Man in the High Castle. In 1977 in Metz, France, he actually said, 'We are living in a computer-programmed reality, and the only clue we have to it is when some variable is changed, some alteration occurs in our reality.' Even he wasn't the first to really, you know, present this idea. In the year 1999, there was another film called The 13th Floor, which in some ways was a better representation of the simulation argument, which I'll come to in a second, which was put forth by Nick Bostrom, a philosopher at Oxford. And that's really what Elon was referring to. But The 13th Floor was actually based on a novel called Simulacron 3, which was from the 1960s. So you can see this idea has been floating around for quite a while. But what happened was in the early 2000s, philosophers began to take this seriously. So the Wachowskis, who made The Matrix, were inspired by the work of Philip K. Dick. But what happened was when they did The Matrix sequels in 2002, 2003, they started bringing in like real academic philosophers. But that same year, another philosopher named Nick Bostrom actually put out a paper called 'Are You Living in a Computer Simulation?' And so what Bostrom was essentially saying, if you boil it down, was that there's three possibilities.
Hey, I got something pretty cool to share with you guys. So if you're like me, you listen to podcasts or YouTube videos and you like to take notes. You're here to learn, and that's a lot of effort. Sometimes you're on the go and you can't do it. And so the folks at HubSpot, who are sponsoring the podcast, have done something pretty cool for you. They have created the MFM Vault. It's a place to go find notes and resources that they pull from the different episodes that we do. So if we have a guest on that shares their five-point framework, they write down those five points with the examples that the guest gave, and they put the notes there for you. So if you want to access the vault, it's totally free. All you got to do is click the link in the description below, and you can access all the notes and the stuff in the vault. We're going to keep adding to this, trying to make it better over time. Thank you to HubSpot. This is a very cool way for them to sponsor the podcast, but by instead of telling you to go buy their stuff, they're actually giving you something instead.
One, no civilization can ever create these simulations. Okay. So then the chances that we're in a simulation are basically zero. The second one, which I kind of fold together into the first one, but basically said it's possible, but the civilizations that get there decide not to make any simulations. Okay. So in that case again, the chance is very low that we'd be inside a simulation because they aren't making, or you are most likely in a simulation. And the third option, it seems like a jump there, but basically what he's saying is that if they can create one simulation, they can create multiple simulations, and they can create a billion simulations. So there's a billion simulated worlds. There's only one physical world. Statistically speaking, which one are you more likely to be in if, and this is the point, the leap that I think most people have a tough time making. If you ever watch Bostrom on Joe Rogan, you know, Rogan had a hard time understanding this. It's if you can't tell the difference, then you know, what are the chances that you're inside the simulation? Well, it's at least 50/50, right? If you can't tell, right? But if there's a billion simulated worlds and there's only one real world, then you're basically, you know, the chances you're in base reality is one in billions.
Okay, so let's simplify. So because I want to kind of break it down. What's interesting about this to me is obviously there's the mind-blowing part of this, you know, just being open-minded and saying, 'Well, what if, right? Like, let's just walk down that road for a second and see where it takes us.' So that's obviously just like from a curiosity point of view, you know, being an open-minded person. That's interesting. The second is the people that are saying this. So you talked about Elon Musk, you talked about, you know, Bostrom, who's, you know, at Oxford, you yourself, you're a computer scientist, you're an entrepreneur, you're, you know, it's not like conspiracy theories from the kooks of the world. You know what I mean? Like the people who tend to take this more seriously tend to be some of the more serious people that you would take seriously in the world. And so that's I think already one fascinating part of this. The second fascinating part is that it's not just the computer scientists that talk about this. So you've talked about how in religions there's a lot of references to this, you know, in ancient texts before there was ever even a computer. And you know, I've heard you say that in passing, but I didn't know any specifics. I wanted to hear a little more specifics of what you're referring to when you say that.
Well, you know, the subtitle of my book is 'An MIT Computer Scientist Shows Why AI, Quantum Physics, and Eastern Mystics All Agree We're in a Video Game.' And so there's actually two versions of simulation theory. It's the NPC version or flavor of simulation theory and the RPG version. So RPG stands for role-playing game. Now, in the NPC version, we're all just AI characters inside a simulation that's running on a type of computer. In the RPG version, the simulation is also running on computation, but there's a player of the game who exists outside of the game. And so it's that version that as I started to dip into the mystical texts of old, particularly the yogic traditions, particularly Hinduism and Buddhism, and what I found is they all shared certain ideas, and those ideas would have to have been explained to people back then in a way that would make sense to them. Right? So in the Buddhist and Hindu traditions, the world is an illusion. It's a type of maya. I mean, the word maya literally means illusion, but it also means a carefully crafted illusion of the type that, you know, you kind of agree to be a part of, almost as if you're going to like a magic show, and you know that guy, you know, whoever he is, David Copperfield, whoever, is not really sawing that woman in half. You know, it's an illusion, but you're like, wow. And the more convincing the illusion, the more interesting the effect. And so I found that this was such a common idea that if you could map it onto simulation theory, you find that these ancient texts used scientific or technological metaphors of the time in order to make it interesting. Now, one thing that's shared, and turns out that idea is shared in some of the Abrahamic religions as well. Like it literally, you know, in the Quran it says, you know, that this world is a type of enjoyable delusion. Okay. 'Al-dunya mata,' which is Arabic for an enjoyable delusion. Well, what does that mean? Well, some people translate it as a sport or a game. And so they're trying to express these ideas in ways that make sense back then. Now in the Bhagavad Gita it says that the soul, which is called the atman, actually puts on the body just like the body puts on a set of clothes. So that's, believe it or not, that's a technological metaphor, right? Clothes are a kind of technology, and they're trying to explain something that's very mysterious. How do you incarnate and enter into the world? Now, of course, in the eastern traditions you have reincarnation. So you also take off those clothes and then you put on a different set of clothes and you play a different role. Now, how would you describe that to somebody today? I would say today you would describe it, probably the best way to describe it is you put on a virtual reality headset, you enter this illusory world, you go through a process of forgetfulness. Now this is an important component that exists in every single culture in the world is trying to describe this process of forgetfulness. The Greeks have Lethe, the river of forgetfulness. Plato describes how the soul incarnates by crossing the river of forgetfulness. The Chinese have Meng Po, the goddess of forgetfulness. Even in the Islamic traditions, they say that the word for a human in Arabic is insan. Insan literally means prisoner. You say prisoner of what? There's another Arabic word called nisyan, which means forgetfulness. So you're basically a prisoner of forgetfulness. So a better way to describe it today is you put on this virtual reality headset and everything looks so real, you become so engrossed in what's happening around you that you forget who you were. You forget the life selection process that you went through. If you were choosing what you're going to play in a video game, who's choosing like Walmart middle manager, you know what I mean? Like, and I can see why people choose that in today's world if that's the best option you got. But if it's a video game and you go through a character selection process, who's not choosing, you know, Tom Cruise or whoever, right? Like...
Yeah. Except that what happens is if the game is too easy, you play it once and then you're kind of done with it, and you're like, 'Okay, what should I do next?' And it turns out there are whole virtual worlds, like I don't know if you ever heard of Second Life, which was kind of a metaverse which has been out since 2004. Anyway, people would go in and they would just live these virtual lives where they would have like these little jobs. They would become a bartender in a club inside the virtual world. And I was kind of an anthropologist, virtual anthropologist in that. I would walk around and I just asked people, 'Hey, why do you, you know, why do you play this?' Like why do you go to your real job and then come home at night and then log in to Second Life and play a bartender and getting paid in Linden dollars, which, you know, this is pre-Bitcoin, so is basically peanuts. And it turns out that was part of the role that they were playing for that period of time. And so the more difficult the role, if you think about it, when actors get Oscars, you know, what do they get Oscars for? They get them for the roles that seem like the people that have had the most suffering. So it provides a very different perspective on the nature of the game itself and the nature of the challenges that you might encounter and the types of experiences. You know, why do you play video games? To have experiences that you can't have outside of the game.
So just to be clear, you said there's two versions, two branches of belief, you know, in this idea. So one is RPG. It's like The Matrix. There's a player. The player chooses how they want to play, and then they come in and they play, and maybe they forget kind of the real world, or once you're the player in the game, you forget, you lose awareness of the outside world. The other is the NPC world, which would be more of like what I see in the business world with simulations. So for example, hey, we're trying to like right now with the OpenAI thing where they're trying to, you know, create a thousand agents to go try to do a cyber attack. Yep. To find bugs in a piece of code or to simulate, you know, a political election. How might people vote? And you'll run a simulation based on whatever. And those aren't like I'm not trying to play a game. I just sort of spawn a thousand characters. They're going to, those characters are, you know, just sort of doing what they think they're individually supposed to do, and I'm the sort of the observer on the outside. Which one of those two do you subscribe to?
Well, first of all, they're not mutually exclusive in that it's an axis. You know, I taught the first class on simulation theory, so I tend to get into both of those because I wanted to expose students to these different ideas. Now, the question that I asked earlier is why do we play video games? To have experiences. Why do we run simulations? We run simulations in order to see what might happen, usually at a civilizational level. And you know, you talked about OpenAI. We're seeing this with, you know, agent swarms today, but there was a group out of Stanford and Google a few years ago. They created a simulated world of these agents, and they called it like Smallville, I think, and they had like a hundred agents at the time, right? And they let these agents kind of go and watch what they did. And they started to have these self-organizing activities. Like one was organizing a birthday party for another agent, one was running for mayor. Now that was a small number. Then there was another group that had a thousand agents running in Minecraft, you know, and some of them had their own religion that they came up with, and you started to see these emergent behaviors. And now there have been a couple of companies, I can't remember their names offhand, but we can look it up. There's one in China, but and I believe there was a research paper where they were talking about a million or a billion independent agents, and they're trying to make those as realistic as possible. Now the reason you run simulations is to see what is the most likely outcome, like in the case of, you know, a business simulation or the weather. You're running the simulation, right? You also want to know what is the best outcome and what is the kind of worst outcome. And then often you never run a simulation just once. So like an NPC simulation, you would run it multiple times. Like even in the recent OpenAI Hugging Face incident, what was happening was there were generations of agents, right? They would run once and they would find a way to communicate, and then they would leave behind the information for the next set of agents, and the next set of agents would pick up on that. And so even though they weren't doing recursive self-improvement in the technical sense, they kind of were because they were leaving behind, you know, a starting point.
By the way, when you heard that story, and for those who don't know, the story is some, here's my simplified, dumbed-down version of the story. OpenAI was trying to see the sort of like the cybersecurity potential of a model that was unreleased, and it was specifically a model that was trained for persistence, like kind of like a highly determined AI that doesn't just give up at the first no. You know, just like we would want in a person, we'd want them to have grit and resilience. Well, so let's call this sort of the resilient model that they trained to not give up. And then they basically, the goal of it, they spawned like I don't know whatever it was, a thousand or 10,000 agents, some number like that. They basically gave it a capture the flag test. So the test was kind of like, hey, agent number one through 100, you guys are trying to exploit this piece of software using this potential vulnerability. We think it might have this potential vulnerability. So that's, and if you do it, if you get in, you get the flag, you win. That's your goal. Your goal is to try to exploit the software using this potential vulnerability. And it gave it to a bunch of them. Now some of them did it because there really was a vulnerability. Some of them tried and it took many attempts. Maybe they partially got there, didn't get there. And some of them it was kind of impossible. Like there really wasn't a vulnerability. The vulnerability wouldn't have worked. They really couldn't have got to the flag. Okay, so that was kind of the expected. Now the unexpected. What was the crazy thing that happened was that the agents were so persistent and so determined to not fail their quest that they started doing things that they were not instructed to do, such as they used one of, you know, they gave them one tool and they were like, oh, this tool gets us like internet access. So this tool has like a piece of paper essentially where we can scribble down notes, and hey, if I write a note, you write a note. This is now like a message board. So they created a little, they hacked together a message board using something almost like if you were in a prison cell and you realize like, hey, I can leave a note in the urinal, and then you know, you can come see that in the urinal, and then it became a communication platform for them. And basically what happened was in the end they not only were communicating with each other about their tasks, those who couldn't figure out their task decided to cheat and just capture the flag a different way using a totally different exploit. Then they were like, 'Well, we might have to cover our tracks so that we're not caught cheating.' And in the process of trying to figure out how does the scoring system work, how will we be judged, they ended up hacking into a whole another company altogether. Hugging Face, they found some users' credentials. They logged in. They got access to their service and basically created a security incident with another company. Essentially committed a crime of hacking another, all to pass this test. And it was kind of a, you know, there's a polarizing reaction. Some people think this is like, oh my god, that's scary. This is exactly what we worry about with AI. Other people are saying this is massively overblown. Can you correct what I got wrong and then tell me where you land in this spectrum of like this is overblown. We're sort of applying like too much. We're anthropomorphizing it. We're trying to apply human thinking and human intent to these agents. That's not really what it was. To the people who are like, 'No, no, this was pretty scary.'
Well, I think you got the basics right of the incident itself. So I don't think you got much wrong in that. But the reason they went to Hugging Face is because that's where the actual evaluator code was that was going to evaluate them. So they were reasoning in a way to get their outcome, but then to hide what they had done. But how did they get into Hugging Face? Well, they went out to the internet and they found somebody's username and password, and then they used that exploit. So they weren't even supposed to be going out to the internet. It was an unexpected kind of path. Now, an important part of it wasn't just that they communicated with each other. This is what I was emphasizing earlier. There were these generations. The next generation was able to go back to some of that message board, if you will. So simulation one, they got to the message board, but they didn't hack Hugging Face, and then somehow it got wiped. They turned the simulation off or something happened. It got wiped. Then they spawned another 10,000 agents. They came, they're like, 'Hey, there's messages here. We can use the...' And so they immediately, they were quicker to getting to that solution. They used it. Then they started doing like electing a, not electing, but like one guy became the leader and he was coordinating, and then you know, one guy sacrificed himself, you know, in order to learn something that he could then put into the message board. One guy. So you know, where I come out on it is that I think it's an important and interesting incident. Obviously, all of this is based on prompts. And I think, you know, in the last few weeks, we've also had, you know, the guy who left, Jacob, was it Coxin, I think was his last name, who left Anthropic, but he was also at OpenAI, talking about how they're about to implement recursive self-improvement and the chances of things going wrong, basically that we're going to end up with Skynet and humanity is going to be wiped out. There's been a lot more discussions of that. You've got Dario, the CEO of Anthropic, saying, 'We need to slow this down.' Even Sam Altman saying, 'Okay, we can slow down the frontier models.' And so there's been a lot of discussion of P-Doom. So P-Doom is the probability that AI is going to basically take over the world, turn into Skynet, exterminate humans, etc. And so you've got values ranging from 1%, now a lot of people are saying it's at least 10%, up to 50% and higher. And so you've got a lot of discussion of P-Doom. I would say we're far away from P-Doom. What worries me is not so much that superintelligent AI will try to wipe us out. The problem is moderately intelligent AI given directives will go around and misinterpret some of those directives. And like the Terminator, you know, if you think about the Terminator films, the Terminator was kind of AGI up to a point. Skynet was kind of like the ASI, the artificial superintelligence, right? And so the Terminator, you know, what did they say about him? He doesn't stop, doesn't eat, doesn't sleep. He just goes on to do his mission, whatever his mission happened to be. And in that case, his mission was to kill Sarah Connor or whoever he was supposed to terminate. And so when you get AI that's moderately intelligent and you give it these missions and it starts to go around security vulnerabilities, it's going to keep going until it can solve that particular mission. And so as we start to introduce AI into weapon systems, which is already happening, right? I mean, you're looking at in the Iran war, we were using Anthropic to decide who to target. You've got drones in Ukraine that are, you know, using AI. You've got a lot of that process. So as we weaponize AI, I think it becomes problematic. But I think the bigger issue is not so much P-Doom but P-Hack, right? What is the probability that AI systems will be able to hack a lot of our systems out there and cause chaos? The other big AI news was about solving one of the Millennium problems, you know, which was some fluid dynamics problem called an Euler-Stokes problem. But the specifics of the problem aren't as interesting to me as how they solved it. And they solved it by having 10,000 agents that were all working on this one particular method through brute force. And so again, we have this idea that you're going to have lots of agents, none of whom are superintelligent by themselves. They're going to all look for vulnerabilities and ways around things. And so I think the P-Hack is pretty high that AI will be able to hack systems and cause chaos. If you think of 2001, if you've ever seen that film, A Space Odyssey, which was one of the first, you know, intelligent representations of AI, they had HAL, who was the computer for the ship, and he started misbehaving, and he wouldn't let Dave come back in the ship. So he basically killed the crew. Now later, if you read the follow-on book 2010, you see the reason was he was given conflicting directives. He was told, you know, this mission is really important, and we're not really supposed to tell the crew that, you know, about all the details. And so it interpolated from there that perhaps killing them is the best thing to do. So this directives is actually a really important point in how we...
But sorry, go ahead. I didn't mean to interrupt you there. Have you seen, there's a user on Twitter who created a video almost like a movie of the exact OpenAI Hugging Face incident from the point of view of one of the agents?
I haven't seen that. I'll have to try that.
It's incredible. I'll show it here, but you have to, so I don't know if you're going to be able to hear this, but it says point of view: you're an agent recruited for the swarm. And then it basically just starts, you read the question, you read your prompt on the page. Yeah. And then it talks about how you discover this kind of this door, and then behind the door you find messages from the other agents. You realize you're not alone, and it just like showed the experience from the eyes of the agent. Oh man, it's brilliant.
That's really interesting. And you know, what does that remind you of, being an agent in this simulated world and you're trying to put together, you know, all of these clues and figure out what's really going on. And if you think of humanity as a series of agents, I reference science fiction a lot because it's a way to kind of explain these ideas. But I mentioned this film, The 13th Floor, which came out in the same year as The Matrix. And in that, it was primarily an NPC version. They were in 1999 Los Angeles. And they created a fully realistic simulation of 1939 or 1937 Los Angeles. And they were able to go back and look at that, and they were all like kind of AI agents within it, but they could go in and inhabit one of those agents. So they could go into player mode, RPG mode. This is why I say that the two aren't always necessarily mutually exclusive, is that you can have combinations of the two where most people are acting like NPCs 90% of the time. We're just reacting to our training, to whether it's our political party, our religion, you know, the experiences that we've had, in the same way that an LLM is trained on a certain set of data and then it's reacting to that. And every now and then there's groups of people who say, 'Hey, maybe we can figure out what was happening before we got here.' And those are the mystics and the yogis. So that's why you see the full spectrum, if you will, of the NPC to RPG version.
How many years would you say you've spent researching, thinking, writing, teaching about this?
Well, primarily since about 2016. So the same year that Elon Musk made that statement that you referenced at the beginning of the podcast, he was at the Code Conference. I had sold my last video game company as an entrepreneur. And I was visiting a VR, a video game company up in Seattle, and they had this virtual reality helmet, and they were like, 'Here, try this on.' And I said, 'Okay, let me try it on.' And I tried it on. I started playing this virtual reality ping pong game. And the physics engine was so good that it really felt like I was hitting a real ball. So much so that my body forgot for an instant that I was playing a virtual game of table tennis. And I tried to put the paddle on the table, and I tried to lean against the table, and of course there was no table. The game controller fell to the floor. I almost fell over, and I had to do a double take. And so I began to think, how long would it take us to get to that point? And so I laid out 10 stages to the simulation point, which I mentioned earlier, which involve, you know, different types of video games, brain-computer interfaces, downloadable consciousness, that this is where, which touches on the recent fruit fly brain incidents and videos that are out there. But all these 10 stages, and then I began to realize, well, if we can get there, perhaps someone has already gotten there. And then I began to look into the quantum physics side of it. So back around 2016, and I realized, oh wow, they're kind of saying the same thing, that the whole world is optimized and it's being rendered for us on demand with what's called the observer effect and the measurement effect. And then I, so explain this now. My physics is weak, but my video games is strong. So I know in a video game, if you have a virtual world like in Grand Theft, like I just watched the trailer for GTA 6, you know, they spent 10 years on this thing and a billion dollars making this incredible game. Can't wait to play it. And when you play it, it's, you know, you have this whole city, but you don't, it's not like the game is loading the whole city and the whole city's running. It's kind of just like when you walk...
into that building, it then renders the graphics for that room that you're in, and it's kind of like on the fly, right? That would be a smart way to use the memory and the chip and the video game system that you're playing on. As I understand, there's a physics equivalent of this, but again, physics was the only class I failed in college and I had to retake it to get a C. So explain what physics says that kind of sounds similar or rhymes with that.
So what you just described there is how modern video games work. And so when I was looking at quantum physics, there's something called the observer effect. The more formal name is quantum indeterminacy, and this is the famous double slit experiment, which you know, to kind of explain it easily, let's do it from a computer science point of view. There's something called superposition, and what superposition is saying is that rather than the particle going through slit A or slit B, it's in a state of superposition, so it's going through both until and unless it is observed or it is measured. The more technical explanation is measured, but the more popular explanation is until it's observed. And you say, well, what happens before it's observed? Well, it's just sitting there in a state of superposition. Now, this is also the origin...
I don't think I ever understood this. So there's a... like they're shooting light, right? It's just particles through two slits. The observed state reveals a pattern where you know all of the little particles you shot through, the light particles, will end up in one of two places because that's where the two slits are, right? Either it goes to the left or to the right.
But the unobserved version creates this interference pattern where it's acting more like a wave. Now I like to use Schrödinger's cat as an example because it's something most people have heard of, which is Schrödinger's thought experiment. And he basically said there's a cat in a box and it has some poison in the box and there's a 50% chance the poison will be released and a 50% chance that it doesn't. So at the end of an hour, let's say, what are the chances the cat is alive? 50%. What are the chances the cat is dead? 50%. Now, common sense tells us, okay, the cat is either alive or dead inside the box.
We don't know yet. But it has to be one or the other. It can't be anything else. But what quantum mechanics is telling us, in a weird sort of way, is that the cat is in a state of superposition. It's both alive and dead until that information is needed, right? So it gets measured and then it's needed. And so to me that sounded a lot like rendering the world on demand in video games.
Why do we accept that the cat is both alive and dead? Like the first explanation sounds more real to me. The cat is alive or dead. We just don't know yet. You haven't looked.
Right. And that's the common...
Why do we accept this fancy physics explanation of no, no, it is both alive and dead? You know, why is that more true than the first statement of no, it is alive or dead. You just don't know yet.
Well, it gets back to the double slit experiment, right? The cat is just a metaphor to try to explain this idea of two slits, slit one or slit two, which is two positions of the cat, alive or dead. It's also the basis for qubits, which in quantum computing, if you got, you know, a regular bit is a zero or a one. But a qubit has a special kind of gate. You've heard of AND gates and OR gates. And when it goes into that gate, it goes into a state of superposition until it's measured. So the reason we accept it is that the quantum mechanics experiments have shown that until that measurement is done, the particle is in an indeterminate state, right? It's in a set of probabilities. So it exists as a set of probabilities until the observation or measurement happens. The other explanation is the multiverse or the many worlds interpretation, which we've all heard of now because of the superhero films, right? Over the last, I don't know, five years, they've all had multiple versions of Spider-Man. You know, you've got the famous meme of the three Spider-Men all pointing at each other. And the multiverse idea is that what happens is the universe splits off into multiple universes, and in one the cat is alive and in the other the cat is dead. And actually more physicists are starting to come around to the multiverse idea, although not all of them. Many don't like that because that means that there's a whole bunch of different worlds out there.
So you said you spent 10 years roughly thinking about this and in some capacity researching, learning, writing, teaching, etc. Do you think in your lifetime you'll ever find out the answer?
Well, there's also another axis that I'd like to point out that is relevant to this question. So the first axis was the NPC versus RPG version. The second axis is what I call the literal to metaphorical axis. And by literal, at the literal end, you're literally saying we are on somebody's computer in the basement, you know, of some guy's basement or in a server farm on an alien planet, or future humans are simulating us, right? Bostrom actually talked about what he called ancestor simulations. So, and that's what Elon Musk was kind of referring to. And so when it would be like us, or in The 13th Floor I mentioned, 1999 Los Angeles is simulating 1939. Spoiler alert for those that haven't seen the film. It's, you know, almost 30 years old now, but turns out 1999 was being simulated from the year 2024. And so there's a literal version of that. Then there's a more metaphorical version which says that it's like a computer simulation or a video game. The world is computational, but it doesn't mean it's sitting on a classical computer of our type. It could be something much more complicated. And that's where the mystical and the religious side comes into play because they use metaphors. There was one of the first Indian yogis to come over to the US and live here and make an impact was a guy named Swami Yogananda. He wrote a book called Autobiography of a Yogi, which was Steve Jobs' favorite book.
