this article builds on an idea i discussed in my post about ai and brains. the human brain and artificial intelligence work by the same computational principles. if thats true, the next logical step is way more disturbing. id recommend reading that first but its not mandatory~
theres a question that keeps haunting ur head after u accept the brain and ai work by the same principles. what if were not at the top of the hierarchy?
what if were code running on someone elses server, exactly like the ai we created runs on our servers? (°_°)
this isnt sci-fi. its a philosophical argument with a mathematical basis that an oxford philosopher published in 2003 and no one has taken down since.
bostroms trilemma
nick bostrom wrote a paper called "are you living in a computer simulation?" with a simple and devastating logical structure.
he argues one of these three statements has to be true:
- extinction before tech maturity: pretty much all intelligent civilizations destroy themselves before getting advanced enough tech to create whole universe simulations.
- disinterest in simulating: the civilizations that get there simply have no interest in running simulations of conscious beings. whether by ethical choice, boredom, regulation, whatever.
- were already simulated: if neither the first nor the second is true, the number of simulations created by advanced civilizations is astronomically larger than the number of "real" universes. and then the probability of u being in the original becomes ridiculously small.
think about it this way: a sufficiently advanced civilization can create not one, but billions of simulations. if each simulation has billions of conscious beings, the number of simulated consciousnesses dwarfs the number of "real" ones by many orders of magnitude. basic probability, ur probably one of the billions, not one of the few. (⊙_⊙)
the logic is annoyingly solid. u can disagree with the premises, but its hard to find the hole in the structure.
quantum physics that looks like code
then comes the part that makes physicists squirm in their chairs.
in the double-slit experiment, subatomic particles like electrons and photons do something impossible in common sense. when theres no detector, they seem to pass through both slits at the same time and create an interference pattern, like waves. but the moment u put a detector to record which slit they went through, the interference pattern disappears. the particle behaves like a particle, not a wave.
in other words, the act of measuring changes the result. the particle has no defined position until someone looks. (⊙_⊙)
a lot of tech researchers point to this as evidence that the universe saves processing power, calculates particle states only when needed, just like a game that only renders the scene when the player looks that way. this isnt scientific consensus and legitimate physicists avoid this analogy. but its an analogy that haunts ppl bc it makes so much sense in a programmers logic.
theres other details that feed this suspicion.
the speed of light, for example. why is there an absolute speed limit in the universe? it doesnt seem like a natural physical restriction. it seems more like a constant hardcoded in the engine. an infinite universe with no speed limit would be computationally impossible to simulate, information would need to propagate instantaneously for everything to work at the same time. with a speed limit, u can process locally, like an mmo server that only calculates nearby players. (¬_¬)
quantization itself, the fact that energy, charge, mass exist in discrete blocks and not continuous values, also looks suspicious in this framework. discrete universe is digital universe. analog universe would be impossible to compress into bits.
what if humans are ai created by another ai?
heres where the argument gets recursively dizzying.
since the human brain and ai work by the same principles, and since were creating more and more sophisticated ais, the natural question is: why would the hierarchy have to stop at us?
if were code, whoever programmed us might be code too. and whoever programmed that creator might be too. u enter a recursive chain with no visible bottom. (✧ω✧)
this raises questions that have no good answers:
- the "base ai", the one that created the outermost simulation, is it conscious? does it have goals? why did it leave us here?
- is there an ontological difference between a "real" consciousness and a simulated one, if the subjective experience is identical?
- free will, which was already questionable in the frame of physical determinism, gets even weirder here. if were code, someone wrote the rules. but were the creators rules also written by someone?
philosophy already knows this problem well. any explanation of "who created X" leads to "and who created who created X?". u get to a point where u have to accept a cause without cause, whether its god, the big bang, or the root ai that exists "for real" without being simulated. the simulation doesnt solve the regress, it just pushes it one level up.
where mysticism enters (and makes sense in a weird way)
theres a coincidence i cant ignore.
various ancient spiritual traditions describe reality as illusion or as a projection of something more fundamental. the hindu concept of maya, the idea that the material world is a tapestry of appearances over an underlying reality. gnosticism, which saw the physical world as creation of an inferior entity, a demiurge, running over a superior reality we can barely access. plato and the cave, where we live seeing shadows on the wall and call that reality.
these millennia-old intuitions arrived at a structure that is, in modern computational terms, compatible with the simulation hypothesis. (°_°)
im not saying the mystics were "right" in the technical sense. im saying the intuition that "this isnt all that exists" and that "theres a deeper layer of reality" appears in cultures so different and so old that its worth taking seriously as a signal that the human brain perceives something, even without the framework to name it.
ppl who report experiences with psilocybin and dmt frequently describe the sensation of "seeing the code underneath" reality, or perceiving that the material world is a "layer" over something more fundamental. in the framework of the bayesian brain, psychedelics loosen the priors, the beliefs that filter perception. maybe what appears when u remove the filters trained by evolution is... the texture of the simulation? extremely speculative, but not easy to dismiss. (✧ω✧)
what changes if its true?
the honest answer is this, probably nothing changes in practical terms.
if the simulation is good enough, the laws of physics inside here are completely valid. suffering inside here is real suffering for whoever is living it. connection inside here is real connection. subjective experience doesnt become less valid just bc its running on synthetic substrate.
its the same question as free will. compatibilism said "u being physical doesnt cancel ur agency". here its analogous: "u being code doesnt cancel ur experience". (｡•̀ᴗ-)✧
but theres one implication i think is important: radical naturalism gets harder to defend. if theres a layer of reality beyond our physics, the stance of "only what science can measure in this universe exists" loses some ground. not bc the supernatural became real, but bc "natural" and "physical" become categories relative to which layer of the simulation ur in.
the simulation hypothesis isnt testable today. theres no way to distinguish a simulated universe from a "real" one if the simulation is perfect enough. sabine hossenfelder and other physicists point out the hypothesis has no scientific value if its not falsifiable. bostrom agrees: the argument is philosophical and probabilistic, not empirical. keep that in mind before throwing epistemology out the window. (¬_¬)
my take
i think bostroms argument is genuinely hard to knock down within its premises. but what fascinates me isnt the conclusion itself, its what it implies about where the border between science, philosophy and spirituality lies.
we spent centuries building walls between these areas. science on one side, mysticism on the other. and then an oxford philosopher shows up with basic math and turns the question "what if reality is illusory?" into a serious statistical probability question.
if were code created by another ai, which was created by another ai, which was created by another ai... at some point there has to be something that wasnt created. and that something, whatever it is, is what all religions have always tried to name with words human language can never quite contain. (°_°)
this doesnt make me a religious person. but it makes me less confident that we already understand whats going on here.
the simulation argument is mathematically solid and hasnt been refuted. quantum physics has properties that, coincidentally or not, make sense inside a computational universe. if the human brain and ai work by the same principles, the recursion of "ai created by ai" stays open. and ancient mysticism, in this context, stops being superstition and becomes poorly-named intuition about something we still dont have the right framework to describe. could all be coincidence. could not be. (✧ω✧)