- Virk, Rizwan.
The Simulation Hypothesis.
Cambridge, MA: Bayview Books, 2019.
ISBN 978-0-9830569-0-4.
-
Before electronic computers had actually been built,
Alan Turing
mathematically proved a fundamental and profound property of
them which has been exploited in innumerable ways as they
developed and became central to many of our technologies and
social interactions. A computer of sufficient complexity, which
is, in fact, not very complex at all, can simulate any
other computer or, in fact, any deterministic physical
process whatsoever, as long as it is understood sufficiently to
model in computer code and the system being modelled does not
exceed the capacity of the computer—or the patience of the
person running the simulation. Indeed, some of the first
applications of computers were in modelling physical processes
such as the flight of ballistic projectiles and the
hydrodynamics of explosions. Today, computer modelling and
simulation have become integral to the design process for
everything from high-performance aircraft to toys, and many
commonplace objects in the modern world could not have been
designed without the aid of computer modelling. It certainly
changed my life.
Almost as soon as there were computers, programmers realised
that their ability to simulate, well…anything
made them formidable engines for playing games. Computer gaming
was originally mostly a furtive and disreputable activity,
perpetrated by gnome-like programmers on the graveyard shift
while the computer was idle, having finished the
“serious” work paid for by unimaginative customers
(who actually rose before the crack of noon!). But as the
microelectronics revolution slashed the size and price of
computers to something individuals could afford for their own
use (or, according to the computer Puritans of the previous
generations, abuse), computer gaming came into its own. Some
modern
computer games have production and promotion budgets larger
than Hollywood movies, and their characters and story lines have
entered the popular culture. As computer power has grown
exponentially, games have progressed from tic-tac-toe, through
text-based adventures, simple icon character video games, to
realistic three dimensional simulated worlds in which the players
explore a huge world, interact with other human players and
non-player characters (endowed with their own rudimentary
artificial intelligence) within the game, and in some games and
simulated worlds, have the ability to extend the simulation by
building their own objects with which others can interact. If
your last experience with computer games was the Colossal Cave
Adventure or Pac-Man, try a modern game or virtual
world—you may be amazed.
Computer simulations on affordable hardware are already
beginning to approach the limits of human visual resolution,
perception of smooth motion, and audio bandwidth and
localisation, and some procedurally-generated game worlds are
larger than a human can explore in a million lifetimes.
Computer power is forecast to continue to grow exponentially for
the foreseeable future and, in the Roaring Twenties, permit
solving a number of problems through “brute
force”—simply throwing computing power and massive
data storage capacity at them without any deeper fundamental
understanding of the problem. Progress in the last decade in
areas such as speech recognition, autonomous vehicles, and
games such as Go are precursors to what will be possible
in the next.
This raises the question of how far it can go—can computer
simulations actually approach the complexity of the real world,
with characters within the simulation experiencing lives as rich
and complex as our own and, perhaps, not even suspect they're
living in a simulation? And then, we must inevitably speculate
whether we are living in a simulation, created by
beings at an outer level (perhaps themselves many levels deep in
a tree of simulations which may not even have a top level).
There are many reasons to suspect that we are living in a
simulation; for many years I have said it's “more likely
than not”, and others, ranging from Stephen Hawking to
Elon Musk and Scott Adams, have shared my suspicion. The
argument is very simple.
First of all, will we eventually build computers sufficiently
powerful to provide an authentic simulated world to conscious
beings living within it? There is no reason to doubt that we
will: no law of physics prevents us from increasing the power of
our computers by at least a factor of a trillion from those of
today, and the lesson of technological progress has been that
technologies usually converge upon their physical limits and
new markets emerge as they do, using their
capabilities and funding further development. Continued growth in
computing power at the rate of the last fifty years should begin
to make such simulations possible some time between 2030 and the
end of this century.
So, when we have the computing power, will we use it to build
these simulations? Of course we will! We have been
building simulations to observe their behaviour and interact
with them, for ludic and other purposes, ever since the first
primitive computers were built. The market for games has only
grown as they have become more complex and realistic. Imagine
what if will be like when anybody can create a whole
society—a whole universe—then let it run to
see what happens, or enter it and experience it first-hand.
History will become an experimental science. What
would have happened if the Roman empire had discovered the
electromagnetic telegraph? Let's see!—and while we're at
it, run a thousand simulations with slightly different initial
conditions and compare them.
Finally, if we can create these simulations which are so
realistic the characters within them perceive them as their real
world, why should we dare such non-Copernican arrogance as to
assume we're at the top level and not ourselves within a
simulation? I believe we shouldn't, and to me the argument that
clinches it is what I call the “branching factor”.
Just as we will eventually, indeed, I'd say, inevitably, create
simulations as rich as our own world, so will the beings within
them create their own. Certainly, once we can, we'll create
many, many simulations: as many or more as there are running copies of
present-day video games, and the beings in those simulations
will as well. But if each simulation creates its own
simulations in a number (the branching factor) even a
tiny bit larger than one, there will be exponentially
more observers in these layers on layers of simulations than at
the top level. And, consequently, as non-privileged observers
according to the
Copernican
Principle, it is not just more likely than not, but
overwhelmingly probable that we are living in a simulation.
The author of this book, founder of
Play Labs @ MIT,
a start-up accelerator which works in conjunction with the
MIT Game Lab,
and producer of a number of video games, has come to the same
conclusion, and presents the case for the simulation hypothesis
from three perspectives: computer science, physics, and the
unexplained (mysticism, esoteric traditions, and those enduring
phenomena and little details which don't make any sense when
viewed from the conventional perspective but may seem perfectly
reasonable once we accept we're characters in somebody else's
simulation).
Computer Science. The development of computer games is
sketched from their origins to today's three-dimensional
photorealistic multiplayer environments into the future, where
virtual reality mediated by goggles, gloves, and crude haptic
interfaces will give way to direct neural interfaces to the
brain. This may seem icky and implausible, but so were pierced
lips, eyebrows, and tongues when I was growing up, and now I see
them everywhere, without the benefit of directly jacking in to a
world larger, more flexible, and more interesting than
the dingy one we inhabit. This is sketched in eleven steps, the
last of which is the Simulation Point, where we achieve the
ability to create simulations which “are virtually
indistinguishable from a base physical reality.” He
describes, “The Great Simulation is a video game that is
so real because it is based upon incredibly sophisticated
models and rendering techniques that are beamed directly into
the mind of the players, and the actions of artificially
generated consciousness are indistinguishable from real
players.” He identifies nine technical hurdles which
must be overcome in order to arrive at the Simulation Point.
Some, such as simulating a sufficiently large world and
number of players, are challenging but straightforward
scaling up of things we're already doing, which will become
possible as computer power increases. Others, such as
rendering completely realistic objects and incorporating
physical sensations, exist in crude form today but will
require major improvements we don't yet know how to
build, while technologies such as interacting directly with
the human brain and mind and endowing non-player characters
within the simulation with consciousness and human-level
intelligence have yet to be invented.
Physics. There are a number of aspects of the physical
universe, most revealed as we have observed at very
small and very large scales, and at speeds and time intervals
far removed from those with which we and our ancestors
evolved, that appear counterintuitive if not bizarre
to our expectations from everyday life. We can express them
precisely in our equations of quantum mechanics, special
and general relativity, electrodynamics, and the
standard models of particle physics and cosmology, and
make predictions which accurately describe our observations,
but when we try to understand what is really going on or
why it works that way, it often seems puzzling and
sometimes downright weird.
But as the author points out, when you view these aspects of
the physical universe through the eyes of a computer game
designer or builder of computer models of complex physical
systems, they look oddly familiar. Here is how I expressed
it thirteen years ago in my 2006 review of Leonard Susskind's
The Cosmic Landscape:
What would we expect to see if we inhabited a simulation? Well,
there would probably be a discrete time step and granularity in
position fixed by the time and position resolution of the
simulation—check, and check: the Planck time and distance
appear to behave this way in our universe. There would probably
be an absolute speed limit to constrain the extent we could
directly explore and impose a locality constraint on propagating
updates throughout the simulation—check: speed of light.
There would be a limit on the extent of the universe we could
observe—check: the Hubble radius is an absolute horizon we
cannot penetrate, and the last scattering surface of the cosmic
background radiation limits electromagnetic observation to a
still smaller radius. There would be a limit on the accuracy of
physical measurements due to the finite precision of the
computation in the simulation—check: Heisenberg
uncertainty principle—and, as in games, randomness would
be used as a fudge when precision limits were hit—check:
quantum mechanics.
Indeed, these curious physical phenomena begin to look
precisely like the kinds of optimisations game and simulation
designers employ to cope with the limited computer power
at their disposal. The author notes, “Quantum
Indeterminacy, a fundamental principle of the material
world, sounds remarkably similar to optimizations made
in the world of computer graphics and video games, which
are rendered on individual machines (computers or mobile
phones) but which have conscious players controlling and
observing the action.”
One of the key tricks in complex video games is
“conditional rendering”: you don't generate the
images or worry about the physics of objects which the player
can't see from their current location. This is remarkably like
quantum mechanics, where the act of observation reduces the
state vector to a discrete measurement and collapses its complex
extent in space and time into a known value. In video games,
you only need to evaluate when somebody's looking. Quantum
mechanics is largely encapsulated in the tweet by Aatish Bhatia,
“Don't look: waves. Look: particles.” It seems our
universe works the same way. Curious, isn't it?
Similarly, games and simulations exploit discreteness and
locality to reduce the amount of computation they must
perform. The world is approximated by a grid, and actions
in one place can only affect neighbours and propagate at a
limited speed. This is precisely what we see in field
theories and relativity, where actions are local and no
influence can propagate faster than the speed of light.
The unexplained. Many esoteric and mystic traditions,
especially those of the East such as Hinduism
and Buddhism, describe the world as something like a dream,
in which we act and our actions affect our permanent
identity in subsequent lives. Western traditions, including
the Abrahamic religions, see life in this world as a temporary
thing, where our acts will be judged by a God who is outside
the world. These beliefs come naturally to humans, and
while there is little or no evidence for them in
conventional science, it is safe to say that far more
people believe and have believed these things and have
structured their lives accordingly than those who have adopted
the strictly rationalistic viewpoint one might deduce from
deterministic, reductionist science.
And yet, once again, in video games we see the emergence of a
model which is entirely compatible with these ancient
traditions. Characters live multiple lives, and their actions
in the game cause changes in a state (“karma”) which
is recorded outside the game and affects what they can do. They
complete quests, which affect their karma and capabilities, and
upon completing a quest, they may graduate (be reincarnated) into
a new life (level), in which they retain their karma from
previous lives. Just as players who exist outside the game can
affect events and characters within it, various traditions
describe actors outside the natural universe (hence
“supernatural”) such as gods, angels, demons, and
spirits of the departed, interacting with people within the
universe and occasionally causing physical manifestations
(miracles, apparitions, hauntings, UFOs, etc.). And perhaps the
simulation hypothesis can even explain absence of evidence: the
sky in a video game may contain a multitude of stars and
galaxies, but that doesn't mean each is populated by its own
video game universe filled with characters playing the same
game. No, it's just scenery, there to be admired but with which
you can't interact. Maybe that's why we've never detected
signals from an alien civilisation: the stars are just procedurally
generated scenery to make our telescopic views more interesting.
The author concludes with a summary of the evidence we may
be living in a simulation and the objection of sceptics
(such that a computer as large and complicated as the
universe would be required to simulate a universe). He
suggests experiments which might detect the granularity
of the simulation and provide concrete evidence the
universe is not the continuum most of science has assumed it
to be. A final chapter presents speculations as to who
might be running the simulation, what their motives might
be for doing so, and the nature of beings within the
simulation. I'm cautious of delusions of grandeur in
making such guesses. I'll bet we're a science fair project,
and I'll further bet that within a century we'll be creating
a multitude of simulated universes for our own science
fair projects.
January 2020 