How much of reality does a simulation actually need to simulate?
I’ve been thinking about the arguments that the universe is simply too information-rich to simulate, and I’m not sure I buy the premise.
Sometimes when I listen to music I’m completely under its spell. I hear the song as a single object. Other times I have what I call “fresh ears” and I suddenly hear how the trick is done—the individual instruments, an imperfection in the voice, the recording, the edits. Same information coming in, but I’m interrogating it at a different level.
Science seems to do this to reality. We keep developing fresh ears and looking underneath the illusion: molecules, atoms, nuclei, quarks; better telescopes, better microscopes, more sensitive detectors.
But here’s the thing: discovering quarks doesn’t mean a simulation now has to render every proton in the universe down to its quarks.
Same with JWST. It could look almost anywhere, but it doesn’t. We’ve made some finite set of observations. The simulation only has to generate those observations and make them consistent with everything we’ve already seen.
So maybe the computational requirement doesn’t scale with the information content of the apparent universe at all. Maybe it scales with the amount of information we’ve actually forced reality to reveal and commit to.
What I’m trying to figure out is whether there’s a way for observers to defeat that strategy. Can we interrogate reality in a way that eventually forces a lazy simulation to actually maintain the underlying world?
Or can it always just render one step ahead of us?