25 Years Later, The Thirteenth Floor shows the Simulation Hypothesis Better than The Matrix
By Rizwan Virk
Warning: Spoilers for both films, The Matrix and The Thirteenth Floor ahead!
This year is the 25th anniversary of both The Matrix (see my article here) and The Thirteenth Floor (released on May 28, 1999). Both films depict what we now call the simulation hypothesis, the idea that we might live inside a computer simulation. In my college-level class that I teach about the emerging field of simulation (titled Simulation Theory: Sci-Fi, Technology, Religion and Philosophy), while The Matrix is the first sci-fi film I assign to my students, the second one is The Thirteenth Floor. While The Matrix is by far the most recognizable popular media depiction of the simulation idea, on this anniversary I am arguing that The Thirteenth Floor may be a better and richer representation of a number of aspects of the simulation hypothesis than even The Matrix.
While almost everyone I know has seen The Matrix, many fewer have seen The Thirteenth Floor, which was based on Daniel Galouye’s 1964 sci-fi novel, Simulacron-3. While neither film was highly anticipated (the most anticipated film of 1999 was definitely Star Wars Episode I: The Phantom Menace), The Matrix became the breakout hit and has moved into history as the film to fuel a thousand late night philosophical discussions about the nature of reality. Which means The Thirteenth Floor to become a niche, cult classic with those obsessed with simulation theory (like me!). While both movies present important aspects of the simulation idea, if you were to ask which one “does it better?”, I would give a slight preference to The Thirteenth Floor.
Before we dive into the reasons why, I think it’s important to explain the distinction between the NPC vs. RPG flavors of the simulation idea. NPC stands for non-player (or more formally, non-playable) characters, and in this version, the characters in the simulation are all AI’s driven by code and do not have any independent existence outside the simulation. The term NPC caught on first in single player video games for all the people that you don’t control in the game — the bartender, the merchant, the farmer, the opponents, even the princess you might be rescuing. In the RPG version, which stands for role-playing game, the players exist outside of the simulation and have graphical characters, called “avatars”, who live within the simulation and which we control. In this sense, the RPG version is like playing today’s video games. It’s important to note that these two aren’t mutually exclusive — you can have NPCs and avatars inside the same multi-player video game like World of Warcraft or League of Legends or Fortnite. In fact, the two flavors are more like the ends of an axis — on the one end are simulations where all the characters are AI, and on the other end are video games where 100% of the characters are avatars controlled by players.
The Matrix didn’t delve into the distinctions between these poles as much as the emerging field of simulation theory needs. The NPCs were mostly, using 1990s terminology, “programs” like Agent Smith and the Oracle (among others) and all the main characters, including Neo, Morpheus and Trinity, existed outside of the Matrix and through BCIs (Brain Computer Interfaces), controlled characters inside the simulation. This discussion about sentient AI centered around the AI that had taken over the world and was enslaving humanity, a sci-fi trope that is still being discussed today (perhaps more seriously than it was in 1999).
The Thirteenth Floor, on the other hand, provides a rich backdrop for the discussion of sentient AI of various types within ancestor simulations. An ancestor simulation, a term coined by Oxford philosopher Nick Bostrom in his landmark 2003 paper about simulation theory, Are You Living in a Computer Simulation?, means a simulation of the past (of our ancestors, in the general sense of the word). If, for example, we were to do a computer simulation of ancient Rome, that would be considered an ancestor simulation. The simulation in The Matrix is a kind of ancestor simulation, as it simulated the Earth at the end of the twentieth century, while the real year in the “real world” is much later (Morpheus explains that no one really know what year it is). Technically, since the matrix was built by the AI that had taken over the world in the early twenty first century, you could say that the humans were the ancestors of the AI. However, in The Matrix, this ancestor simulation was to keep the humans busy while they were being farmed by technology; it wasn’t to really to simulate the past per se, and its focus was primarily an RPG simulation — meant for the players.
In The Thirteenth Floor we see two representations of ancestor simulations which are closer to the idea that Bostrom had in his original paper (which, incidentally led to the term the simulation hypothesis). At the start of the film, we are in Los Angeles in the year 1999 and in one of the labs, on the 13th floor, there is a secret project that has created a holograph-like simulation of Los Angeles in the year 1937. This is a better representation of an ancestor simulation because the characters in 1937 are primarily NPCs — they are simulated AI characters who live, work, love and sleep within the simulated world, going about their lives in the same way that we might go about our lives today, blissfully unaware that they might be in a simulation.
One of the first ideas that’s brought up about this ancestor simulation is: are the people in the simulation real? Are they leading “real” lives? Are they living conscious lives? This touches on the bigger question of free will vs. determinism. To be fair, the NPCs (terminology that I’ll use even though it doesn’t appear in the film) in the simulation are just AI, so from the perspective of the outside world, they could be considered to not be real, not conscious and not have free will. Yet when Douglas Hall, the main character, enters into this world, he finds the world so convincing that he has a hard time accepting that they are “just” NPCs. In fact, at one point, when he comes back to the “real” world he declares, “They are just as real as you and me!” They get up in the morning, they go to work, they have affairs, they worry about each other, and they have secrets, just like us.
This tension of the reality of the lives of the AI characters is one of the central points of the film, as is the case of whether they have free will. The mechanism that is used to enter the simulation is a little less specific than in The Matrix, which “plugged in” through a port at the back of the neck, what we would call a Brain Computer Interface today. In The Thirteenth Floor, a character gets into a pod, is scanned by a laser, and magically finds themselves inhabiting the lives of the one of the characters in the ancestor simulation. There is a twist here, and that is that you end up in the character that was modeled after you, so like in The Matrix, your avatar is what I like to call identomorphic — looks and feels exactly like you — except that you don’t have the memories of the NPC, and while you inhabit it you are free to do what you would like. In one interesting plot twist, Douglas is investing the death of Hannon Fuller, his mentor and the gentleman who invented the simulation, and finds that he has taken over his character and went to a hotel to meet various young women (AI Girlfriends and Boyfriend are already here btw, minus all the sensations that you can get an ancestor simulation).The NPC version of Fuller then doesn’t remember when it was taken over by the player.
This gives us an interesting twist on the NPC vs. RPG narrative, and a new way to view free will. You could argue that the NPC is deterministic — it runs based on code. You could also argue that when it is inhabited by a player, it is no longer an NPC, it is an avatar of a person outside the simulation, and the “will” element comes from the player. This echoes our own on-going debate on the nature of consciousness between materialist scientists and religious adherents who believe that there is a soul, that exists outside of the “fake world”. The Thirteenth Floor provides us with an interesting alternative — where free will both does and does not exist. Rather than being always in one or the other, we can be both: in “NPC mode” we are basically running off of our programming. In “avatar mode, we are in direct contact with our “player” (soul) or consciousness outside of the simulation.
Speaking of the hotel where Fuller hangs out, the bartender, Ashton (who helped Fuller make arrangements with the girls) plays a key role in the film. While both are sci-fi, the genre of The Matrix was an action film, while The Thirteenth Floor is more film noir, with a murder (Fuller’s) which drives Douglas Hall into the sim in the first place. While investigating the murder of Fuller ( his mentor), Douglas realizes that Fuller has left a message for him in the only place that is safe: inside the simulation. Douglas realizes that the message has been left with the bartender.
We have only one problem, the bartender, who seems as real as any bartender Douglas has encountered in the real world, turned out to be curious enough to not only read the message but investigate it. The message is about the world not being real; he drove out to the edges of LA and realize that the world they inhabit (the 1937 version of LA) … the roads and buildings turn into green gridlines we associate with the scaffolding of a virtual world or video game. This turns his world upside down … and he asks a relevant question to anyone who has contemplated being inside a simulation: “Why would you want to do this to us? Why are you fucking with our minds?” This is a good question and applies to an NPC simulation — with AI that feels emotions, experiences happiness and sadness, triumph and suffering. It is also a question that many have asked of our creator, if there is one.
The film’s noir theme wouldn’t be complete without a mysterious woman who claims to be Fuller’s daughter, Jane, who shows up in the real world (in 1999) to collect her inheritance and shut down the simulation. Since Fuller was Douglas’ mentor, and he had never met (or heard Fuller talk about) a daughter before, this adds to the mystery woman, with whom Douglass ends up having an affair with before she disappears. Add a curious detective and the neo-noir genre atmosphere is complete. Douglass tracks the woman down only to find out that she is a grocery clerk who has no memory of him, nor does she claim to be Fuller’s daughter.
This of course brings up the big plot twist, when Hall figures out, through Jane that she is from “outside of this world”. In other words, she is inhabiting the grocery clerk character as her avatar. This is the same thing Hall did to his character when he enters the ancestor sim. She informs him that 1999 is not the real world either. It is an ancestor simulation along the same lines as the 1937 sim! The real year is … wait for it … 2024! In the real world of 2024, she is in fact Fuller’s daughter in “the real world” and Douglas is an NPC based on her real world husband, with whom she is no longer enamored because of his cruelty not only towards her but to their creations.
Douglas Hall now has an existential crisis similar to the bartender in the world “below”. In an interesting plot twist (which mirrors the Matrix trilogy), the bartender is able to “upload” out of the simulation into his “player”. This brings us deeper into the issue of whether, even if we are AI, we could ever “escape the simulation” — a popular topic in simulation theory.
There are some experiments going on now about whether we are in a simulation, and rumors of billionaires funding experiments to see if we can escape the simulation, and well respected AI researchers looking at how AI might escape as a template or framework to think of how we might escape our simulation.
Finally, you can see that we are bringing up the idea of stacked or nested simulation. In The Thirteenth Floor, Jane tells Douglas that in the real world, they created thousands of simulations, but out of all of those, only one version (i.e. this version of 1999) was the only one that went on to create its own nested ancestor simulation. Therefore, they must shut down the simulation because it is now using more computing power than is allocated.
Many opponents of simulation theory use the argument of nested simulations to say that if we are in a simulation, then it must be “simulations all the way down”, which of course isn’t really possible. In The Thirteenth Floor, we see how the dilemma might be handled: if we are in a simulation, and we create more simulations, and they were to create more simulations, this would quickly use up the computing power from base reality and the simulations would have to be shut down, just as Jane was going to do to Hall’s version of 1999.
This brings us to the bigger question of the purpose(s) of the simulation, if we are in one, and why it might be important to make sure our simulation is not shut down. In 2019, a philosophy professor, Preston Greene, wrote an op-ed for the New York Times saying that we shouldn’t try to find out if we are in a simulation, because that might cause the simulators to shut us down. It would put us at a risk of termination. This was a technological version, in my opinion of Pascal’s Wager about God — if we act like there is a God (rather than acting like there isn’t), we are better off. If God exists, we’re better off because we followed his commandments and end up in heaven; if he doesn’t’ exist, it won’t matter — we’re not worse off than if we acted like he doesn’t. On the other hand, if he does exist and we act like he doesn’t, we could end up in Hell, a much worse outcome.
In the novel that The Thirteenth Floor was based on, Simulacron-3, there is also a termination risk. The purpose of the simulation in the novel is to test out products (think Mad Men or advertising executives on steroids) — basically a souped-up market research tool or focus group that runs on a computer with a large number of artificial beings or NPCs. In the world of Simulacron-3, it wouldn’t do to have the participants in your experiment know they were in an experiment, because it might skew the results. In the film, the simulation of 1999 (i.e. Douglas Hall’s world) was being shut down because they had created a nested simulation.
Which brings. us to the the final twist, where Hall is able to use the same technique to “escape” the simulation and take over his player’s body in the year 2024.
All of this brings up perhaps one of the most important questions relating to simulation theory. Should we even try to test and find out for certain if we are in a simulation? And in a related question, should we try to build highly realistic nested simulations as a way to test if we are in a simulation?
The Thirteenth Floor, now celebrating its 25th anniversary, tells us not only to be careful with playing with the emotions of our AI creations, but to be careful about creating nested ultra-realistic simulations that might tax the simulation’s computational power. If we do, the simulators might cause us to suffer the fate of Douglas Hall’s version of 1999: being shut down.
And unlike Hall, we might not be able to so easily escape to the “real” 2024.
Note: Rizwan Virk was the founder of Play Labs @ MIT and author of The Simulation Hypothesis: An MIT Computer Scientist Shows Why AI, Quantum Physics and Eastern Mystics Agree We Are in a Video Game and The Simulated Multiverse. He is currently at Arizona State University’s College of Global Futures in the Center for Science and the Imagination, and visiting The Leverhulme Center for the Future of Intelligence at the University of Cambridge.
Follow him on X @rizstanford, on Instagram @rizcambridge and at zenentrepreneur.com.
