Are you Living in a Me-Simulation?
In Yesterday’s Post I tried to explain why I thought I was living in a computer simulation, and in particular a first-person simulation where I would be the only full-time-conscious being.
Unfortunately, I wanted to explain too many things at the same time, without taking the time to give precise arguments for any of them. I ended up spending only one sentence on why I believe I am living in such a simulation, and did not even try to give a precise estimation of the probability of being in such a simulation.
Luckily, Ikaxas took the time to understand what were the premises, and gave precise counter-arguments. It forced me to be much more precise, and updated my beliefs on why I thought I was living in a first-person simulation. I am grateful he did so, and will be re-using arguments we exchanged in Yesterday’s discussion.
The Simulation Argument
In what follows, I will assume the reader is familiar with the simulation argument. I will be referring more precisely to this text: https://www.simulation-argument.com/simulation.html.
Ancestor simulations are defined in the simulation argument as “A single such a computer could simulate the entire mental history of humankind (call this an ancestor-simulation) [...].”
The simulation argument estimates the probability of ancestor simulations, and not first-person simulations. Therefore, I cannot infer the probability of a first-person simulation from the probability of being in case 3) of the simulation argument (almost certain to live in a computer simulation), as I did in Yesterday’s post.
However, Bostrom mentions the case of what he calls selective simulations (Part VI., Paragraph 13), and in particular me-simulations:
In addition to ancestor-simulations, one may also consider the possibility of more selective simulations that include only a small group of humans or a single individual. The rest of humanity would then be zombies or “shadow-people” – humans simulated only at a level sufficient for the fully simulated people not to notice anything suspicious. It is not clear how much cheaper shadow-people would be to simulate than real people. It is not even obvious that it is possible for an entity to behave indistinguishably from a real human and yet lack conscious experience. Even if there are such selective simulations, you should not think that you are in one of them unless you think they are much more numerous than complete simulations. There would have to be about 100 billion times as many “me-simulations” (simulations of the life of only a single mind) as there are ancestor-simulations in order for most simulated persons to be in me-simulations.
In short, me-simulations are simulations containing only one conscious being and one should only believe that he lives in a me-simulation if he thinks those are (at least) 100 billion times more probable than ancestor-simulations.
I briefly introduced at the beginning the concept of first-person simulations. I define first-person simulations as a particular case of me-simulation, where only the perceived environment of the conscious being is rendered (like it is the case in a first-person shooter).
I will try to explain why I think the probability of being in a me-simulation is comparable to the one of being in an-ancestor simulation.
Incentive and Probabilities
In this part, I will suppose that we are not in case 1) of the simulation argument (where “the human species is very likely to go extinct before reaching a posthuman stage”), and try to understand what would motivate a posthuman civilization to run ancestor simulations and me-simulations.
In the next paragraphs, I will assume that the reader is familiar with the Fermi Paradox and the concept of Great Filters. If this is not the case, here are introductions by Kurzgesact:
Fermi Paradox: https://www.youtube.com/watch?v=sNhhvQGsMEc
Great Filters: https://www.youtube.com/watch?v=UjtOGPJ0URM
Now, my claim is that one of the following statement if true:
i) Posthumans will run ancestor-simulations because they don’t observe other forms of (conscious) intelligence in the universe (Fermi Paradox), and are therefore trying to understand if there was indeed a Great Filter before them.
ii) Posthumans will observe that other forms of conscious intelligences are abundant and the universe, and will have low interest in running Ancestor simulations.
The premise of this claim is that human-like civilizations able to invent computers/internet/spaceships are either extremely rare (Great Filter behind us), or they are not, and in this case there exists a Great Filter ahead of us (because of the Fermi Paradox).
So Posthumans will only have interest in running Ancestor-simulations if they believe their form of intelligence is extremely rare. In this case, they would want to run a large amount of those simulations in order to have precise-enough estimations of how likely is each event in their past (e.g. Cambrian explosion) and determine which events were extremely unlikely.
As a consequence, we would be living in an ancestor-simulation if and only if all of the following statements are true:
1) There are no Great Filters between us and the Posthuman running the simulation
2) There is at least one Great Filter behind us
3) The simulation passed this first Great Filter, and his now “playing” the digital age phase (1990-2018)
Because of the Alignment problem, and other existential risks in general, I personally find the statement 1) extremely unlikely (less than one in a thousand).
Now, assuming the statement 2) being true, the probability of being in a simulation which successfully passed at least one Great Filter and is now full of intelligent species capable of exploring space (in SpaceX I believe) is extremely unlikely. In fact, the ratio (number of simulations that passed a Great Filter F) / (number of simulations that arrive to the Great Filter F) is by definition small (smaller than 0.000001 in my intuition of a Great Filter).
Therefore, I believe that the probability of being in a 2018-Earth-like Ancestor-simulation consisting of 7 billion humans is extremely small (less than a one-in-a-billion chance).
Physicalism and Consciousness
Before going into more details with Me-Simulations, I need to make sure I was clear enough about how I see Ancestor-Simulations in contrast with Me-Simulations, and in particular how they differ in the consciousness of their inhabitants.
Physicalism essentially states that “everything is physical” and that there is nothing magical about consciousness. Hence, if posthumans want to simulate human civilizations with as a starting point the Big Bang, they must implement enough complexity in whatever makes the simulation work so that the humans in the simulation are complex enough to (for instance) build spaceships, and from this complexity will naturally arise consciousness.
I will not discuss here the probability of Physicalism being true in my model of reality. For this post, I assume Physicalism to be true in order to reason conveniently about consciousness of simulated minds (I don’t see how we could discuss consciousness in simulations with Dualism for instance).
However, I will answer one issue Ikaxas addressed in Yesterday’s post:
“[...] even in a first-person simulation, the people you were interacting with would be conscious as long as they were within your frame of awareness (otherwise the simulation couldn’t be accurate), it’s just that they would blink out of existence once they left your frame of awareness.”
I claim that even though they would be conscious in the “frame of awareness”, they would not be fully-conscious. The reason is that if you give them sporadic consciousness, they would greatly lack of what constitutes conscious-human’s subjective experiences: identity and continuity of consciousness.
I have close to none personal empirical evidence about the existence of Me-Simulations. The closest phenomenon of first-person simulation (particular case of me-simulation) to my personal experience would be Virtual Reality (VR).
To compare the probability of Me-Simulations I will start by explaining why I am convinced that Me-Simulations can be made cost-efficient, and then I would discuss the usages I believe I posthuman civilization could have of me-simulations.
Virtual Reality and Complexity
I grew up playing video-games where only a tiny fraction of the fictional universe is rendered at each moment. In 2017 I went to a VR conference where they explained how 2017 would be a VR winter because plenty of startups would be developing VR Software but there would be neither investment nor public interest. Moreover, the complexity of rendering 360 high-resolution images would be overwhelming considering the current (2017) algorithms/computation power.
In particular, I went to see a startup developing an experience where humans could go to the movie… using VR. You would then be virtually sitting on a chair, watching a screen included in the image from your VR headset. Even in this simple environment, the startup founders had to only render the chairs at the left and right of your chair because it would be to computationally expensive to fully render all the movie theater.
I claim that me-simulations could be made at lest 100 billion times less computationally expensive than full simulations. Here are my reasons to believe so:
1) Even though it would be necessary to generate consciousness to mimic human processes, it would only be necessary for the humans you directly interact to, so maybe 10 hours of human consciousness other than yours every day.
2) The physical density needed to simulate a me-simulation would be at most the size of your room (about 20 meters squared times the height of your room). If you are in room it is trivially true, and if you are in the outer world I believe you are less self-aware of the rest of the physical world, so the “complexity of reality” necessary so that you believe the world is real is about the same as if you were in your room. However, Earth’s Surface is about 500 million km squared, so 2.5 * 10^13 times greater. Hence, it would be at least 100 billion times less computationally intensive to run a me-simulation, assuming you would want to simulate at least the same height for the ancestor-simulation.
Furthermore, you would only need to run one ancestor simulation to run an infinitely large number of me-simulations: if you knew about the environment, and had in memory how the conscious humans from the ancestor simulation behaved, you could easily run a me-simulation where a bunch of the other characters are just a copy of what they did in the past, but only one (or a small number of people) is conscious. A bit like in Westworld there are some plots where robots are really convincing, but whenever they are not in the same place (e.g. not in the saloon) their non-consciousness is more apparent.
“Agreed, me-simulations might be much more cost-efficient than an ancestor-simulation, but why would a posthuman civilization want to simulate me-simulations in the first place?” you might say.
Here is a list of scenarios that convince me of the usage of such simulations:
Intergalactic travel : humans traveling at (almost) the speed of light to another galaxy might feel bored and want to relive the life of famous humans. For instance “relive the life of Napoleon and conquer most of Europe in less than 10 years” would be a terrific game to play. In those kind of games you would 1) automatically forget everything about your life (brain wash), before 2) starting from scratch (being a baby Napoleon) the life of Napoleon, until you die and then 3) have your memories back and come back to the spaceship.
Emulated minds: in general, Virtual Reality, as I described before, could lead to me-simulations. Maybe the ping isn’t that great and it is much more convenient to play by yourself (only conscious being) the Napoleon game. Or maybe you are sentenced to live in a virtual world all by yourself. See for instance how minds are transferred into toasters in the “White Christmas” episode of Black Mirror, or even in a Monkey Pillow in the “Black Museum” episode.
Wrapping it up
In the simulation argument Bostrom defines me-simulations which contain only one conscious being.
I defined a particular case of me-simulation (first-person simulation) where only the perceived environment is rendered.
I explained why I believed ancient-simulations where unlikely
I argued that in a first-person simulation (which is for me the most natural way of running me-simulations), only you would be full-time-conscious, the others being partially-conscious without real identity/memories.
I gave arguments in favor of cost-effectiveness of me-simulation compared to ancient-simulations
I gave some intuitive examples of usages of me-simulations for posthuman civilizations