The idea of the world we experience being an illusion or simulation is a very old one, but today I’ll be talking about some of the problems with fictional portrayals of one particular recent variation: immersive virtual reality. These ideas go back to early cyberpunk, more than 30 years ago now, and it’s been 14 years since The Matrix took the idea to bullet-time extremes. So much of this has been discussed before, but it’s still interesting to think about.
Geometry and Geography
Most immersive virtual worlds in fiction have explicitly Euclidean three-dimensional geometries. This is easier to write and to film, although we might except Inception – which had fun playing with toroidal spaces. But real-life computer environments aren’t navigated that way. You accessed the file for this page by clicking on a link or entering an address. You didn’t need to scroll past thousands of virtual books in a virtual library. Even computer games like World of Warcraft and the real virtual world Second Life contain teleportation features, so that players don’t spend forever walking around the map.
The sequels to The Matrix started to address this: there are programmer backdoors that allow a doorway in one virtual room to open into another room hundreds of kilometers away in the simulated space. But that’s pre-specified point-to-point (like clicking on a link). Where is the equivalent to entering an address or running a search routine? Similarly, why were people in Snowcrash willing to ride across an entire simulated planet? Snowcrash was written in 1992, when the web was not yet so big, but Neil Stephenson should still have understood how real-life interfaces are set up. The original TRON (from the early 1980s) is also an offender. Perhaps it can be excused by Kevin Flynn being dropped into a system no one had actually deliberately designed. But in the recent sequel TRON: Legacy, why did he not key the exit procedure to himself instead of to a particular location in virtual space – or at least give himself teleportation? And that leads into the second problem.
Traveling through the Grid in TRON: Legacy. It’s shiny and color-coded for your convenience and features Daft Punk, but why do you need to travel across an entire virtual landscape to use the equivalent of “
Many of the virtual worlds in fictional were more poorly programmed and built than we would expect for someone able to build such a world in the first place. Inception’s dreamscapes run on different rules, but consider Snowcrash. People have fancy head-up displays and optic-nerve interfaces to access their virtual world. Someone deploys a random pattern of noise that, rendered as an image, will be interpreted by the interface in a way as to overload things and cause the user to either be mind-controlled or die (from an ischemic stroke in the visual cortex or something similar). The book ends with Protagonist stopping the virus and replacing the snow crash image with an ad for his antivirus firm, but who designs an interface that could so easily injure the user? Stephenson attempted to explain the virus by saying there was machine code for the human mind and the snow crash accessed it and issued a fatal series of commands. But that’s not how human brains actually work.
Less fatally, there are physics errors in the Matrix: buildings ripple when helicopters hit them, the ground bends when Neo takes off, editing things in the Matrix causes some events from a few seconds before to repeat, shock waves propagate from objects not moving at the speed of sound. And in both TRONs, frisbees fly differently than other objects.
How do you access and leave an immersive virtual world? We’ve talked about the problems with Inception’s interface. Snowcrash has head-up displays to access the Metaverse, which gives vision nicely, but how do avatars receive commands from their users? Especially when those users may be doing stuff in the real world at the same time: walking around or riding a motorcycle? Where is the control loop closed and how does the system distinguish from control motions and simple real-space ones?
In The Matrix, users are connected by plugs at the back of the head, connecting to a complex implanted interface concentrated between the brain’s hemispheres. Good for direct connection to the nervous system, and most of their voluntary nervous systems are redirected into the virtual world, so people who are jacked in don’t move around much. But if dying in the Matrix would kill you in real life, except during the special unplugging procedure or if you’re Neo, why don’t they have a firewall program in place to prevent lethal stimuli from getting through?
TRON has Flynn the elder and the younger both be completely disassembled in real-life and appear in the Grid. This avoids the control-loop problem, but introduces far more serious ones. The simulation now covers not just the data going into and out of a human nervous system. It has to include all of the cells and biochemistry of a human body down to individual neurotransmitter molecules well enough that the in-computer representation can justifiably be called a continuation of Flynn’s person. That is many orders of magnitude beyond all of the computer storage in the world right now (at least six orders of magnitude and as much as twenty-three, depending on the accuracy of the representation), and even further beyond the available processing speed – especially since subjective time in The Grid runs something like 100 times faster than reality. Still worse, the dematerialization and rematerialization system needs to be able to process all of that information and the atoms used to assemble it back into a human body within a few hundred nanoseconds (set by the diffusion times of molecules at 300 K and 1 atm on length scales of synaptic gaps). And in TRON: Legacy, Sam Flynn writes out Quorra, his artificial-intelligence love interest, into a human body. Information and bandwidth aside, where did the material come from? Is there a matter bank somewhere that had held all of the atoms that made up Kevin Flynn’s body and then uses most of them to assemble Quorra? What would happen if you tried to write out someone without having enough material in the matter bank? So TRON doesn’t work, as shiny and awesome–sounding as it may now be.
One last problem with immersive VR: how do you know when you’re not in the simulation? Inception touches on this with its dream-within-dreams and a few other works have addressed it as well. The Matrix’s glitchy physics give glues to people who are still plugged in, but how does Neo know that the next level up isn’t just a better simulation?
For real life, though, this problem is avoidable: unless there should sometime be reliable and sufficiently-otherwise-improbable evidence of being in a simulation (e.g. someone figures out how to hack the location variables and teleports stuff around), we can treat the universe as being a real physical space – although ‘holographic‘ cosmological models may complicate things.