Hide table of contents
You are viewing revision 1.4.0, last edited by Leo

The simulation argument is an argument for the conclusion that, if humanity reaches a stage where it can run sufficiently realistic simulations of its history and decides to run them, we are almost certainly living in one such simulation.

Some philosophers and scientists have argued that it may be possible for sufficiently advanced computer simulations of people to have subjective experiences, just as flesh-and-blood people do. In particular, it might be the case that the experiences of simulated people are so realistic that they are subjectively indistinguishable from those of flesh-and-blood human beings. If so, it wouldn't be possible to tell, from the inside, whether one is a real or a simulated being.

This argument has led some philosophers and scientists to ask whether we could in fact be merely simulated. If so, our future could ultimately be cut short, if the simulation is ever halted.

At least two assumptions are necessary for the hypothesis that we are in a simulation to be possible. First, it must be assumed that simulations would indeed be capable of having subjective experiences. Second, it must be assumed that computers that are powerful enough to run such simulations are technically feasible.

Proceeding from these assumptions, Nick Bostrom (2003) has argued that one of the following three hypotheses must be true:

  1. The human species is very likely to go extinct before reaching a “posthuman” stage
  2. Any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof)
  3. We are almost certainly living in a computer simulation

The core of Bostrom’s argument is that, if simulations are run, then in the long run simulated people will vastly outnumber flesh-and-blood people. He notes, however, that this notion is not necessarily airtight. If the universe is infinite, then statements about the fraction of people who are simulated may be meaningless.

...

(Read more)