The first part of your response was just an assertion that you couldn’t think of a reason to run a universal simulation. I provided one: because starscapes are pretty. Next you’re going to tell me that nobody has ever dabbled in Conway’s game of life, what with it too being completely pointless.
If our universal simulation is just running as basically background noise in the back of the mind of some absurdly large and fast computer (just to keep the circuits warm), then perhaps the computer has the processing power to burn, and is no hurry to get anything done. Possibly it’s reusing code/hardware from a more focused system that actually needed to simulate something resembling subatomic reactions, and what we have now is just the outcome of somebody sparking off a big bang (possibly accidentally) and then leaving everything running on idle.
As I’ve pointed out, reality functions exactly like I’d a stupidly basic simulation to function, if it were writ insanely large. (Something like Conway’s game, specifically.) The thing is, at the lowest levels things are so basic that we can’t rule out the idea that it’s just happening because the fabric of reality just happens to fart quantum fluctuations without anybody making it happen on purpose.
But that’s us. If you woke up as a sentient entity in the world of Wallenstein 3D or Minecraft, you might start to get awfully suspicious about how some of the things around you have rather specific arbitrary behavior for no clear reason - particularly if you’re the only entity around that is demonstrating more than rudimentary intelligence.
And of course if they managed to ‘hack the system’ and observe the underlying memory code they’re operating on, they may become even more certain of an intelligent designer - or, alternatively, they may perceive it the way we perceive molecules and atoms: basic natural law.
I implemented the Game of Life on the PDP-1 on which Spacewar was developed. In assembler.
Sure you can run the simulation as slowly as you wish. You can stop it at any time, you can set breakpoints and look what is going on inside. But if you’ve simulated 15 billion years of the universe at 1% of real speed (which would be really smoking) it takes one and a half trillion years to run. I’m not sure what kind of universe you are positing where the simulation runs, but it would be a lot different from ours. Kind of a long time to wait for pretty pictures, which you can do more efficiently in other ways.
I can absolutely see why a culture would want to do a simulation to experiment with possible universes, and perhaps possible values of constants. It just wouldn’t look from the inside the way our universe looks.
In 1972 when I took AI we were on a path to GAI. We read a collection of papers from 1959 - ditto. Now, every subproblem we studied in the class - chess, doing integral calculus, finding the most efficient routes between places, has been solved. Yet we’re not a lot closer to GAI. I don’t think that’s impossible, just that we’re going about it all wrong.
But, once again, this kind of explanation doesn’t come for free: it brings issues of its own.
For one thing, it’s arguable that it’s not much of a simulation any more – the creator of the simulation is forcing at least some of the results they are going to get.
For another, it affects the probabilities concerned. For example, we can ask what is the chance of living in a “forced” universe vs a unforced one? We don’t know, but arguably the latter would be more likely, since those simulations would be easier to set up.
Finally, if things like the moon can be fake, for the purpose of fooling us, what else might be fake, for whatever motivation? We would have to take it to the logical conclusion that we would have good reason to doubt everything, including your memories up to this moment.
Note I say good reason, because of course we can doubt our reality right now. But that’s a different thing from a hypothesis that posits that most of our objective reality, by volume, is fake. In which case we have much more reason to doubt the remainder.
In a situation where every nearby star has been converted into a Dyson brain, then you’d only need to fake those nearby stars. Including, perhaps the Solar System and all its planets, which may well have been the first system converted.
But another scenario, which I’ve been hinting at occasionally in this thread, is the partial, temporally- and spatially- limited simulation that includes false memories and manufactured input. We might exist as fleeting thoughts in the mind of a Dyson Brain, maybe ten thousand years in our apparent future. If that Dyson brain ever contemplates the distant past, it could fairly quickly whip up an immersive simulation of a single human being, or maybe a population of them, and feed them with sensory input and false memories, to see how they react to hypothetical events.
It is only an expansion of what we do when we imagine other people’s behavior. Right now I’m imagining what my missus will have for breakfast. I know what she likes, and what’s in the fridge. A suitably advanced simulating entity could simulate the fridge, and the local environment, and my missus, and probably get it right more often than I could. You don’t need to simulate Alpha Centauri to run a simulation like that; but you might want to run a fairly detailed simulation of current world politics and events, 'cause I know she’ll be checking out the news on the 'Net just like I am, and that might influence her choice.
Consider a dictionary. No dictionary is smaller than the list of the words it defines. Definitions are simulations of words. Define cat in fewer than 3 letters.
Every item simulated requires structure and instrumentation. Of course the simulation may be virtual and the moon only exists when I look at it. That has serious implications for the simulation techs who are controlling tides. If trees only exist when they are observed, the oxygen/co2 balance is a challenge. The gravitational field measurements would be interesting in a universe where objects only exist when observed.
And then there are those pesky fossils and sedimentation layers in lakes and ice cores and the history carried in our individual DNA. If material, then it all would have to be constructed and maintained at the molecular level. If virtual, it’s existence would have to be calculated with each observation. When I find an arrowhead it’s history and context, is calculated to provide my observation. That’s a lot of instrumentation for each and every movement of each and every simulant.
However, it may be that the simulators are far, far away beyond our 18 billion light year boundary. The simulators may be physically huge with brains the size of the earth and lap-tops the size of Russia. Their enormous sun provides infinite power and their size provides infinite storage. We are just a set for their prime time TV program “Doll House Earth”.
Thanks for the clarification, but we need to move this discussion out of 'simulations;.
I’d like to see another thread discussing the boundaries between GAI, AI, Intelligence, Sentience and Consciousness. Perhaps you could start one in Great Debates. These topics are often discussed on this forum but the conversation is blurred by a lack of definitions.
I’m not sure I follow this point. There’s no obligation to simulate anything.
No apriori reason why a pocket universe needs to have 2 trillion fake galaxies.
Right, and just like the point I am raising about the simulation hypothesis, and Boltzmann brains: assuming the proposition is true, we would have good reason for doubting our reality including the observations that made us think of the hypothesis in the first place. For example, how do we know it is feasible to make lots of simulations? Because silicon-based microprocessors are feasible in our universe. But perhaps the people running the simulation just want us to think simulations are easy to make, for whatever reason?
Who knows what the motives of the people running the simulation might be? Hence, taking the hypothesis seriously, I have (additional) reason for doubting basically everything.
In that case, in my view, you view science as magic with no limits. If we’re on the path to building solar systems, then we’re on the path to building galaxies (what is galaxy except a lot of solar systems?), which means we’re on the path to building superclusters (just a bunch of galaxies), which means we’re on the path to building … and so on. And we’ll just have to agree to disagree.
We know the objects exist, have dimensions and mass, because their interactions in the gravitational field have been confirmed by independent observers over time.
The ‘what if’ scenarios about simulation make the leap from observed science to open speculation… Simulations have properties attributed to Gods. The simulation is omniscient and omnipotent and it is alpha and omega. It is an illusion in the mind of Krishna/Vishnu. Simulation is the just God argument in drag, That is not falsifiable.
In the 400 or so years of scientific observation no evidence of a simulation has been discovered. That may not be a proof but that’s the way you bet.
That’s ridiculous. “We are on a path that, if continued along for billions of years, would lead to enormously large scale endeavors” is an observation. It has absolutely nothing to do with magic. Of course, the energy scale is astronomical, and so are the time scales involved. I find it pretty farfetched that we will last in our current form at our current pace of advancement for even a thousand years, much less a billion. So I don’t think we are gonna be building solar systems. But if our society DID last for a billion years, then even at a GLACIAL pace of technological and population growth, we’d fill up the galaxy in tens of millions of years.
Again, that’s a huge “if”, but the point is that it doesn’t require us to take any crazy turns in the road and develop magical technology that’s indistinguishable from magic. With chemical rockets and solar sails, you could move solar systems – on a time scale of billions of years. So yes, if we continue along our current path for a billion years, that’s where we’d end up.
That you would compare running one human mind to moving a solar system shows that you are the one who believes in some kind of magic, not me – a magic that makes it possible for a human to possess general intelligence at 8 million joules a day, but which would be on part with moving a solar system to duplicate.
Sure, OK.
I mean that’s just speculation, plus my objections to the simulation hypothesis remain, but, yes it’s conceivable for a simulation to include data relayed from the creators’ universe.
For some context, our earliest worm ancestors didn’t evolve until around 555 million years ago, HALF the time scale involved here. 1 billion years is twice the amount of time it took to go from simple eukaryotes to all the variety of life we’ve had since, from you and I to cats and dogs to dinosaurs and birds and fish and lobsters.
Various shortcuts are possible. When you’re not looking, a tree is just part of a simple “general vegetation” calculation that handles the oxygen/co2 balance ,etc.; when you look towards a tree from a distance, it’s just part of the “looks like a forest” calculation. It’s only when you’re actually interacting with the tree that you need to calculate textures, smells and the details of fluttering leaves. The same applies to the Moon.
Duuude, we were just talking about this in Nasty’s dorm room. Ya shoulda been there, it was three am and we were running out of weed when Chucko waltzes in with this huge ziploc straight from Doc’s freezer. It’s this primo Bay Area Bud, and Doc was so wasted he said we could have it if we smoked it all by breakfast time. So all I remember after that was this glimmer of memory where through the fog I see Nads pick his head up off a bean bag chair and rasp “So… what if this is all like… a computer game?” I think everyone said “whoaaaa…” all at once.
Oh, wait, that was 1974… we all recovered from that long ago.
This is analogous to simulating a computer system at different levels of abstraction. You simulate the part you really care about at the transistor level, and other parts at the functional level.
The problem is when the parts interact. Unless the higher level models produce exactly the same results at all times, your simulation will be inaccurate. Plus, you need to keep low level models of everything and swap them in when needed. So it might take less computing power, but it takes more storage.
For the tree example, what happens if a bird lands on your simple tree, eats a bug, and then poops on you, including the bug carcass. You’ll probably never notice, but the people doing the simulation don’t know that. Simpler to simulate everything, especially if another person being simulated in heading for the tree.
Right, and as I say, faking parts of the simulation doesn’t only have practical issues. It also throws up problems like asking “How much of our reality is faked?” and the probability that we’re in a simulation which the whole simulation hypothesis rests on.
That works for 99.9999% of people; but your simulation needs to account for the one crazy entomologist who studies 500,000 bird droppings and meticulously catalogs their contents.
Is your abstracted “average bird diet” algorithm indistinguishable from reality? Even as the environments change, bugs evolve, birds evolve? What if someone tries an experiment where they affect bird diets over an area - does the simulation shortcut account for that?
It better, or scientists would have noticed we are in a simulation sometime in the last 100 years when crunching huge amounts of data became much more common.