The Simulation Argument seems to be resurfacing quite a lot lately, from a talk by Elon Musk, to an article on Singularity Hub.
What is it exactly?
The Simulation Argument was first proposed (to the best of my understanding) by the famous Oxonian philosopher professor Nick Bostrom, and can be found here.
The Argument sates (I am paraphrasing of course) that the reality we are living in may not be as real as we think it is, and this conclusion is reached by making the following philosophical arguments:
We are not the only civilization in existence.
This can be made relying (amongst other things) on the Drake equation, which is a "probabilistic argument used to arrive at an estimate of the number of active, communicative extraterrestrial civilizations in the Milky Way galaxy."
A number of these civilizations have developed the ability to build computer simulations of reality to an extremely high level.
This may be assumed based on Moore's law, which is "the observation that the number of transistors in a dense integrated circuit doubles approximately every two years."
Some of these simulations will be advanced enough to enable the emergence of conscious entities.
We may base this argument on what is called Substrate Independence, which is the assumption that consciousness is not necessarily endemic to Carbon based system, and may also operate on a system based on other substrates (e.g Silicon).
Conclusion
If we consider all of the above to be true, it would make sense that the number of virtual realities exceeds the number of non-virtual realities by orders of magnitude, and therefor the chances are that we are living in a simulation.
Several rebuttals can be found online.
Interesting subject. What about simulations in which simulations are run? Clearly the mathematics of fractals could be profitably brought to bear on the subject. "Simulation," to my mind, is a somewhat fuzzy concept which seems to assume that there is a qualitative difference between a substrate reality and a simulated reality contained therein. From the vantage of someone observing a simulation running within a simulation, no clear criteria seem to distinguish one from the other, beyond the mere given that one contains the other.