A thought experiment..
Concerning the flow of time, humans naturally organize events chronologically—past to present [memory], plans [potential future], and imagination [unassigned].
The relationship between these is how we make sense of entropy or the directional flow of time.
This mental model helps us map reality. If this model breaks down, the result is schizophrenia or clinical insanity. Moving on...
Now, imagine a world with the same temporal properties as ours, where time flows from the past to the future.
Given the three values we've defined—memory [past], plans [future], and imagination [unassigned]—we create agents for this world with these cognitive properties.
Next, we scramble their memories, essentially de-addressing them, and then give these agents instructions to make sense of the relationship between their mental models and external reality.
To make things more interesting, we start with one agent and increase the number by +1 from there. Beyond two agents, whether the number is odd or even becomes very important—you'll see why.
The goal is to observe how an individual agent resolves the conflict between its mental model and external reality in isolation, compared to when it's surrounded by others.
Would the agent accept its mental model as true when alone? And when multiple agents are involved, would they try to use some form of consensus algorithm to resolve the conflict?
How would this resolution occur if their numbers were even?
Would there be separate consensus groups due to the absence of a tiebreaker, or would they reach a unanimous conclusion?
Given the processes employed as the number of agents increases, which consensus would most accurately resolve the conflict?
The answer to understanding time may lie in the consensus mechanism these agents use to resolve the conflict between their mental models of time and external reality.
Doing this with async and parallel computing in Rust & getting my ass kicked cos I don't know what I'm doing.
The idea is simple, give multiple interacting agents files to compress and decompress using different libraries, compare the results based on how they dealt with entropy.
It just occurred to me that I'm building another clock. The consensus mechanism the agents may develop to resolve the conflict between de-addressed memory and external reality [compressing, decompression and entropy] is just a clock, some sort of calendar.
I'm still curious to see how it turns out.
Still working out the kinks...