what about output? and how would the computer simulate the minds and actions of the others the subject interacts with inside of the simulation? when only having to simulate the granularity, the computer would be impinged to have to be able to simulate the other minds and bodies without simulating their actual atomic and molecular processes, etc. this by itself would require tons of computing power as well as information stored about exactly how organisms would behave inside a true realistic system.

this is under the condition that the program was on a computer within true reality and follows the same constraints. perhaps my conclusion is invalid due to the existence of machine/deep learning programming.

can you elaborate what a Spinozian necessity is? if you're saying what i'm thinking, as the subject learns more about his environment and the laws of physics, he would come to conclusions about the laws of physics that, if the computer did not have programmed it in, would lead to a sort of contradiction that the subject would have a hard time reconciling. he would live the rest of the simulation where the law of physics that he has arrived at through mathematics and reason are not congruent with the simulation, and he would always have a suspicion something was going on. because the subject would likely try to pinpoint the law in his environment and as he does, the computer would try to compute something that should exist but doesn't, and this causes the computer/program to break down.