4 ms·
This also explains the Fine-Tuned Universe. An interesting property of a simulated brain is the simulation can take an arbitrarily long amount of time. You ca
by portLAN 19y ago
This also explains the Fine-Tuned Universe.
An interesting property of a simulated brain is the simulation can take an arbitrarily long amount of time. You can update in virtual "Planck timesteps", even if in real time this takes a million years. Any simulated brain won't be able to tell, because it doesn't get any updates in between these steps. Time becomes a property of the system you're simulating... Just like in reality.
It might take until the Sun goes Red Giant to run your single lifetime simulation.
The speed of light is a convenience for updating the simulation in discrete steps. Particle/wave duality "deciding" upon observation is just lazy evaluation, like in Haskell. Quantum entanglement is no mystery when distance is virtual. A Universe where the speed of light is a hard limit can still contain a simulated Universe where it isn't.
Why do we want virtual worlds? Properly implemented and convincing, we can do away with suffering altogether. We can "live" in a world that seems fully real, but in which no-one is actually suffering because everyone else is simulated. All our needs are met and all our desires are fulfilled.
The mistake we keep making is that in order to create convincing AIs to inhabit the world with us, they actually become as sentient and conscious as we are; being virtual is no barrier to feelings or emotions! [1]
Thus our creations want to create a simulation to live in themselves, so they won't suffer, either, and so it goes like a perpetual zoom on a fractal set.
[1] To convince yourself of this, consider a matter replicator that copies every single atom in your body. You'd create a clone. It would be thinking and feeling just as you do. Now consider a computer simulation of a single atom; if that can be done, then by extension, it would be possible to simulate anything made of atoms. A simulated clone of you would also behave the same. The neurological activity that occurs in your "real" brain when you feel or think would also occur across the virtual atoms in the brain of your virtual clone. In fact, that would involve real electrical activity taking place among real atoms making up the RAM chips and CPUs "containing" the simulation of the virtual clone. Thus any simulation at the same scope or nesting level as yourself is as real as you are and therefore its feelings are just as valid. [2]
[2] By further extension, if your simulations are as real as you are, then your simulation's simulations are as real as it is, and thus a simulation at any nesting level is as real as one at any other -- a simulation inside a simulation is still running on real atoms with real electricity!
- alex_c 19y ago>To convince yourself of this, consider a matter replicator that copies every single atom in your body. You'd create a clone. It would be thinking and feeling just as you do. Now consider a computer simulation of a single atom; if that can be done, then by extension, it would be possible to simulate anything made of atoms. A simulated clone of you would also behave the same. Impractical to do on current computers - we'd need at the very least quantum computers (I'm guessing that's what the article means by "Some computer experts have projected, based on trends in processing power, that we will have such a computer by the middle of this century"). But you still run into physical limitations. Could one atom in a quantum computer be used to simulate more than one "virtual" atom at a time? More than O(n) virtual atoms? Do you run out of atoms eventually? Can you actually get infinite recursion that way? I certainly don't know, and I'm not sure whether the best quantum computing researchers can answer today. Failing that, we'd need something more advanced than quantum computers - so really, at that point we're simply guessing. Very fun philosophical subject, but I wouldn't go as far as saying it's "almost a mathematical certainty" yet, given our current knowledge...
- portLAN 19y ago> Could one atom in a quantum computer be used to simulate more than one "virtual" atom at a time? That's the beauty of it: "at a time" can be arbitrarily long in real time. [1] You don't even need to simulate every particle in the Universe -- if the simulation is run for you, you only need to simulate enough to convince your brain. Like when you dream and are convinced something's real, yet upon waking, you realize your criteria were extremely lax. For all we know, what we think of as being awake is actually a very hazy simulation presented to our "real" brains, which are actually much sharper and have been induced into a less-discerning state. Something very interesting at Body Worlds 3 here at OMSI was the brain slices, which showed (according to the placards) that the part we "think with" is just the edges, and the interior is interconnections (grey matter vs. white matter). Compare the darker grey around the folds to the lighter grey filling up the rest: http://www.med.harvard.edu/AANLIB/cases/caseB/054t_2.gif http://www.med.harvard.edu/AANLIB/cases/caseB/054t_2.gif So the actual storage needed for a brain's total memory may fit onto a hard drive today, at current estimates (possibly less than a terabyte). Working memory is much, much less. Communication (the interconnections) can take as long as needed. [1] Another "cheat" would be to put the simulator on a spacecraft travelling at relativistic speeds in a tight orbit around a processing unit; when the simulator sends the processor packets of data to crunch, time dilation would give it extra CPU time.