One thing I never understood about the simulation argument is the idea that simulated people are conscious. This is a necessary component of the argument that there are more consciousnesses inside simulations than outside.
If I run a simulation of the universe on a computer, I don't cause a universe to suddenly exist. A representation of a "person" through variables and symbols doesn't create consciousness. You can see how this is the case by "simulating" a person in your imagination right now. Your mental "simulation" is probably better than we will ever see from a computer program, but you aren't creating a consciousness by imagining things.
Some people will try to rescue the argument by saying that, instead of running the simulation on a computer, future civilizations could find a way to run it on an already conscious brain, like a hallucination. So, they argue that, in the future, the number of total conscious brains that have run these neural simulations will dwarf the number of "non-simulated" people, so you get the probabilistic argument again.
There is a problem with this argument too. I know that whatever my consciousness is, it can be damaged by objects around me. But if the objects around me were simulated, it would not be possible to damage the "host" consciousness using hallucinated objects, just like how someone tripping out on drugs can't give themselves physical brain damage using imaginary things.
Simulation theory advocates could try to punt this by saying that the brain-based simulation would just "simulate" the brain damage. But if they argue that the simulation is running on a human brain just like mine, the argument is equivalent to saying there is a way for my brain to generate and experience the way every possible brain injury would feel without actually sustaining the physical damage, which is absurd and would require a complete revolution in our understanding of the abilities and functions of different parts of the brain.
