What if the universe is pseudo-random?
True Randomness in the universe is widely accepted in physics, and understandably so. In Quantum mechanics, the wave function is thought to be fundamentally probabilistic (meaning the outcome of a wave function collapse is unknowable before it happens), leading many to conclude that the universe is fundamentally probabilistic and non-deterministic. This may be how the universe behaves, but what if that is wrong? What if true randomness is a cleverly hidden illusion?
Computers and pseudo-randomness
Computers are deterministic machines. Given a set of fixed inputs and an algorithm, a computer will generate the same outputs. But what about random number generation? Aren’t random numbers used all the time for everything from video games to secure bank transactions? The trick is that computers are incapable of generating true randomness. Instead, we use pseudo-random number generators (PRNG) to give us sufficiently random-looking data. In statistics, PRNGs are commonly fed a known seed, which is a starting number that will provide deterministic random-looking data (the same “random” data will be generated every time). This allows for computer models that require random data to operate in a testable way while still holding the properties required for random inputs.
In more security-focused applications of PRNGs, we have what are known as Cryptographically Secure PRNG (CSPRNG), which are known to be good if the output of a CSPRNG and “true randomness” are indistinguishable to an observer. These CSPRNGs also have another feature: they are one-way functions, meaning that given a known output, you cannot find the hidden input (the seed). This is known as “one way” because, given a seed and the PRNG algorithm, one can quickly generate the output. However, going the reverse direction from output to known input is impossible (the data can only flow one way). What is unique about CSPRNG and other secure PRNGs is that how the seed is derived is hidden from the user. These tools use sources of entropy to feed the CSPRNG data in a way that is not reproducible to an outside observer. The thing is, though, if you did have a way of understanding how the hidden variables were generated, you could generate the output data in a deterministic way.
Bell’s Theorem, Non-locality, and Bohmian Mechanics
In 1935, Einstein, Podlsky, and Rosen wrote a paper later known as the EPR Paradox. This paper hypothesized that instead of a probabilistic universe, there must be “local hidden variables,” that is, variables unknown to the observer that pre-determine the outcomes. Namely, this was meant to address the behavior of “spooky action at a distance.” The Spooky action is that if two particles are entangled, acting on one particle has an instantaneous effect on the other particle outside of the causal speed limit of the universe (the speed of light). They argued that there must be some hidden “locality” that reconciles this spooky action. While the EPR paper was necessary in pointing out issues in the bounds of quantum mechanics, it was incorrect.
In 1965, John Bell formulated Bell’s Theorem (also known as Bell inequality), a way of addressing the question of whether hidden variables, as proposed by the EPR Paradox, exist. What is fascinating is that Bell’s Theorem was eventually able to be tested in the lab and proved that local hidden variables do not explain spooky action at a distance. How this was interpreted varied widely, and this was seemingly a big win for the probabilistic universe camp. Maybe particles do live in a “superposition” until “observed,” and the wave function collapses in an unknowable and random way. But maybe EPR was only partially wrong. Maybe there are hidden variables, but they aren’t local.
In physics, the locality is the principle that things are only influenced by other things in their surroundings. Locality is bound by the “causal limit” of transmitting information (the speed of light). A unit of information transmitted from one particle to another relies on the causal limits of locality to interact with one another. But Bell’s Theorem proved that particles can (and do) interact non-locally, meaning they interact outside the bounds of the speed of light. The wave function collapse of one entangled particle instantaneously impacts the wave function collapse of another, regardless of distance. Some odd and exciting models could still hold for deterministic behavior: non-local hidden variables and bohmian mechanics.
I love the idea of Bohmian mechanics, also known as pilot wave theory. What I love about this theory is that it preserves discrete particles and velocities, but they are guided by a “pilot wave” that evolves according to quantum mechanics. It relies on non-local hidden variables to reconcile these characteristics and it leads us all the way back to the beginning here: a Bohmian universe is fundamentally deterministic but functionally probabilistic due to the nature of non-local hidden variables.
Does that remind you of anything? It sounds an awful lot like the CSPRNG using a hidden seed and source of entropy. What if our universe is one big one-way function and randomness isn’t really truly random but pseudo-random based on a set of non-local hidden variables and initial conditions hidden from those living inside of the universe?
Now, let’s loop this back to our definition of a CSPRNG. If True randomness is an illusion and instead a cleverly hidden Universal PRNG function itself, then our definition of a CSPRNG should be modified. The output of a CSPRNG is indistinguishable from the Universal PRNG to an outside observer.
Given a set of initial conditions, existence may be deterministic but functionally probabilistic due to the unknowable state of its hidden variables. Maybe, just maybe, we live in a Cryptographically Secure Pseudo-Random number generated universe.
But, of course, that begs the question: where did the seed come from? That’s the topic for another day, but Roger Penrose’s concept of an Aeon in Conformal Cyclic Cosmology offers a pretty compelling seed if you ask me.
Notable Links
We interviewed Stephen Wolfram about ChatGPT (Book Overflow)
LGTM book discussion (Book Overflow)
You might (not) be crazy; your phone (maybe) was listening to you for ads
How am I just learning about racer trash?
Cinamastix on the aesthetics of the “sunlight horror” film Midsommar