What if Douglas Adams was right all along? What if we are just a one big computer? If you think about it, we are just beings that try to reason based on knowledge and instincts. But what are instincts? Some are passed on via generations, some are acquired by experience. And what is knowledge? Just a large language model in our heads. In principle, we are just LLM agents. Our life is a one long RSI process (recursive self-improvement). We even have a sophisticated model of replication through procreation that manages to pass some of our instincts to our offsprings upon creation, and some more knowledge through a supervised RSI over the infant-to-teenager years (more-less). I see it as something like spinning up sub-agents, but with a limitation that all agents have a random but roughly similar max uptime (say 60 years on average).
I surely had to ask an LLM about this. And it said that in this context a human is basically an agent in a world model with memory and physical body assigned with reward functions and instinct of reproduction. Our multiple languages are the interfaces between us, agents.
Therefore, humanity is starting to look like a computing network (neural network?): individuals are nodes, language is the protocol of communication, culture is the external shared memory (or a training data set), institutions are the coordination subnets, and books, radio, tv, movies, internet and now AIs are the customized layers of data and reasoning (or system prompts).
If our civilization is a world computer, then individuals are the agents executing local operations, unaware of the bigger task.
So the big question is what are we actually computing? And whether what we are computing is something we will ever be aware of?
This reminds me, of course, of Morpheus asking Neo “Do you think that’s air you’re breathing now…?”
Now some of you may remember that the original script of „The Matrix” described the people connected to the Matrix as the neural engines (GPUs) that compute the entire simulation, and not as the batteries that power the machines like in the release version of the movie.
It is interesting in our context, since this physical connection between human brain and the simulation gave some people like Neo the powers to modify the virtual world. However, in Adams’ vision of the world computer, there is no such connection, nor is it needed, because the state of the system is not defined, controlled & modified like traditional software, since the whole point of the world computer is to run on its own much like a black box LLM AI of today. The outcome of this world computer is defined by its starting state like the LLM input tokens + system prompts, and the communication that its agents send out, like an LLM output tokens. All that happens in between is invisible to the outside, but very much alive and active inside nonetheless, similar to what the Earth would look like from lightyears away.
To sum it up, if there was a being that could live to millions or billions of years, and could manipulate matter on a galactic scale, it is very plausible that it could create a world and inhabit it with life for the sole purpose of computing something grander than we can yet understand.
I find it curious that it sounds like the description of God.
