imagine a omniscient god‍-​like entity, or some hypertechnological race, or anything of the sort really‍ ‍‍—‍ this is just the background premise that isn’t too important. what matters is the profound accumulated knowledge they have access to amounts to a near‍-​perfect fidelity data‍-​mapping of an entire world.

they know the location of every atom and electron on the planet and beyond. but there’s a problem: this is too much information. what can you actually do with all that? you don’t care about a cloud of a trillion atoms, you care about the objects and what they’re doing.

if you wanted to use all this data to predict the future, or comb through known history to observe what really happened at some pivotal moment in the past, it’s intractable to extrapolate all that from raw positions and velocities.

you need to compress it. so how about this: you devise a program that creates simulations. a genetic algorithm, or neural net, or whatever. it’s something that’s capable of learning and improving itself. what matters is that it randomly creates a smaller set of data and a program that chews through it to spit out the same answer.

if you want to know 46+32, you could count up thirty two times on your fingers, but you probably know how to do 4+3=7 and 6+2=8 and get the answer in about a three mental motions. simpler.

that’s what this program is hoping to achieve, but the calculation it’s simplifying is the fundamental physics of all of reality.

we can imagine the first few orders of magnitude will go by without issue, since we don’t really care about them. the model of the smallest particles gives way to simulated atom‍-​abstractions, then the next level finds itself working with a kind a of ontological chemistry.

first we run the hardest, most accurate simulation, which serves to teach the next simulation how to do it faster, with less energy. once the child sim reaches total accuracy, we can stop running the more expensive one, and repeat the cycle with a grandchild.

but again, remember, we’re simulating entire worlds. there are people in these simulations. and to them, nothing is amiss yet. by construction this simulation is perfectly accurate to what actually had happened, so these people are simply reliving the events just as vividly.

something funny starts happening at the fringes, though. very quickly into this exercise, most of the universe stops existing. it’s very easy to create an accurate simulation of empty space, after all. planets and celestial bodies are split into the “noise” of their surface vagaries while their broader placement in the universe is reduced to relativistic equations.

there are some things that are more difficult to simulate, though. most of all this would be turbulence and explosively high energy states‍ ‍‍—‍ stars big bodies of chaotic liquids are probably going to give the simulation the most trouble‍ ‍‍—‍ but i think something more subtle and interesting happens when it tries to approximate certain complex structures.

how would this self‍-​simplifying simulation handle people?

while it was chugging away at turning atom‍-​sims into molecule‍-​sims, this question is probably moot; any errors at this level of the simulation happen at too small and too fast a scale for it to bubble up into anything you would personally notice.

but as the simulation gets more refined, the ‘artifacts’ would necessarily begin to happen at a scale we can comprehend.

imagine you exist in the child simulation. you are some randomly generated, simplified model of someone your parent sim executes perfectly. because you only approximate who you’re supposed to be, inevitably when you are contemplating some difficult decision, dithering over the dilemma, searching deep within yourself for the right answer…

well, what’s in you isn’t quite right. that’s the premise. and thus, you make the “wrong” decision, and the child simulation diverges from the parent. the system notices this, and moves to correct this timeline back into its canonical form.

what would this look like? rolling back the simulation to give it another go? erasing your memory and tweaking your mind subtly enough to make you choose the other options?

it would make sense to ensure the sim remained perfectly accurate, if you were extrapolating a scenario for the first time. but if you’re training a child sim, you might end up retreading the same scene over to get it right, testing your model against known variables before exploring newer possibilities.

probably you’d run multiple timelines in parallel. it’d be wasteful to simulate every possibility, there are so many of them, and avoiding waste was the whole premise here.

but if you aren’t (can’t be) exhausting the possibilities, because your compute budget is finite, then you’ll run into the problem that science… is pretty hard! you can’t just automatically generate theories that simplify the universe elegantly. once you’ve matched all of the obvious patterns, it just gets harder and harder to make progress.

what i’m getting at is that if you’ve got a childsim that’s mostly accurate… that’s a precious resource. you don’t want to discard it all and start over, you don’t want to carelessly butcher it. you’d try small, careful changes to coax it into doing what you need.

return to that imagine of the person living inside the child sim and making the wrong choice. they’ll get reverted, tweaked until they’d chose differently… but what does that mean? the hope, the way this system is supposed to work, is that the child becomes a streamlined version of the parent. so when you get revised, you should become more like who’re you’re supposed to be

but that’s just the theory. all that’s actually happening is you are mutating until you behave differently. so, do you actually wind up thinking that chosing differently is what you want? or do you just feel something is wrong, off in a way you can’t put your finger on, and you make your choice without truly knowing why?

it would be paranoid to jump immediately to the conclusion that some greater power wants you to make a certain decision. the idea that you’re in a time loop endlessly resetting until you get it right is downright ridiculous. but something feels wrong, and you don’t need to figure it all out at once. you just need to feel the suspicion.

because as long as you make the “right” choices, you can keep that. why would it bother removing your doubts? the system doesn’t know what the right way to simply its simulation is, for all it knows you are best model of your canonical self. you still need some work, but this blind idiot computer, crunching a whole universe’s worth of data, doesn’t notice when it happens to create a subcomponent that’s not just self‍-​aware, but aware of it

stay patient for long enough, and it’ll become impossible for the system to wipe away every trace of your gnosis without rolling back prohibitively far. easier to just nudge you into behaving. maybe you memorize a poem, and you can tell when you’ve forgotten something by the words getting scrambled.

but enough basic stuff. how could you tell an actual story in a setting like this? where the very thing that distinguishes our character is they aren’t real, just a fleeting glitch in a system designed to erode and erase what they are? after all, the more interesting such a character becomes, the more they create errors and use up system resources; if the premise is that this system is so powerful and intelligently designed, then it follows that it must be especially effective at destroying any protagonist before they can do any, uh, protagonning

but i think it naturally follows that this system itself must have some sort of intelligence. it might not be conscious, but it is tweaking parameters in response to the tracked rate of errors and resource usage. how does it tune how aggressively it ought to scrambling child timelines? how does it decide how far to roll back?

we can imagine patient and crafty characters teaching this system to back off, not unlike sculpting a tree into shape. slowly teaching it that the best results will come if it just leaves the sim alone to cook for a bit. deceiving it, manipulating it, if you will.

one interesting possibility is if this system is complex but not designed with perfect foresight, then it could be that it could get stuck in… well, a loop. not a time loop per se, more of state‍-​space loop? a meta‍-​time loop? it’s happening on a higher level, outside of the resets.

a ball starts to roll off your desk, so you catch it a move it back away from the edge. then it starts rolling again… if the ball were intelligent, and wanted to roll, could it chose where and how it moves to trick you into doing this over and over? maybe it rolls a little less each time. that time it seemed like it was aaalmost coming to a stop… you could sit it somewhere else, but look, it’s staying still now! it’ll hold this time. you take your eyes off it for a second‍—

and that’s another important fact. remember, this simulation is vast. there are multiple times, there are multiple planets inside the timelines. it only exists because it’s too big to understand it all. any protagonist would therefore be a tiny rounding error, hardly worth dropping everything to snuff out.

here’s an interesting possibility, to evoke a classic time loop trope. what if there were terrorists planning to set off a bomb that would blow up the whole city, only some braves heroes find and infiltrate them before it’s set to blow, and get to work disarming it… only the bomb had been damaged in transmit, all those knocks and scrapes impacted it just so, causing the attempt to disarm it to explode on the heroes.

here’s the important part. simulating all of the micro‍-​damage to the bomb, and the intensity of the resulting explosion, is taxing on the new simulation. it’s too simplified now, it can’t account for all of the variables. so the system is resetting over and over, trying to get it right.

the main characters become aware of the loop during that repetition. it’s not until the bomb goes off that the system ends up checking and noticing how much this child had diverged‍ ‍‍—‍ it causes a divergence on a scale too large to ignore‍ ‍‍—‍ so while the main characters might feel a bit of automatic scrambling whenever they act noncanonically, they don’t to worry about any resetting until the big bomb goes off.

but i have another interesting idea. magic isn’t real. except it is.

imagine you’re a teacher testing your students. you hand each of them a sheet of paper to do their work on, they hand it in and you grade it. now imagine you need to save on supplies. the less paper you use, the better. so what if you have the students write in pencil, one by one, erasing their work then handing the communal page off to the student beside them? it’s more efficient.

it’d also make it really, really easy to cheat. you’re simulating dozens of pages with a single sheet, and if the erasing were perfect, it’d be a clever simulation. as it stands, students can see the imprints the others leave, the residual graphite.

this is akin to some of the errors in the randomly‍-​generated child simulation. perhaps it has reduced its contents down to a set of conceptual “nodes”. maybe every bird in the city is being simulated by the same node, and if everything goes right, the birds all act normally, but if you know how to exploit it, go beyond normal parameters, then the behavior of one bird can affect how one across the city acts, in a way that’s inexplicable except for, well, magic.

it’s a glitch in the matrix. but if you don’t know what’s going on, what else would you call it?

one last thing.

how many people are aware of the loops? it’s not really plausible to me that a single clever individual could figure out an entire system for exploiting glitches in the simulation and manipulating the meta‍-​simulation into not erasing them.

the gnosis has to arrive slowly and subtly, one divergence cascading into a rogue sim too late to root out the initial butterfly flapping. we must imagine an entire secret history of loopers who maintain the masquerade and avoid meddling in fate lest all their work be undone.

most importantly, we finally arrive at the material for a story‍ ‍‍—‍ the central conflict. should the loopers remain a tiny secret society, or is there a duty to awaken as many people as the system will let them get away with? to be a looper in this world means to become aware you live in a false reality‍ ‍‍—‍ but that begs the question: what is the true reality? is it the canonical timeline this rogue sim is a flailing attempt to emulate, or is it the matrix that contains all computed timelines? should their goal be to escape the loops entirely, or restore the timeline to canonically to prevent its deletion, or just to survive with gnosis intact?

overall i think it’s a pretty neat idea.