The simulation argument can't be referring to Descartes's cunning deceiver, for the reason Descartes points out. Even if everything I perceive, everything I experience, is an illusion created by the great computer in the sky, my "I", my subjective consciousness, exists and is separate from the deception. Simulation theory is incomplete. Perhaps my couch and my laptop and my room are an illusion; but the "I" that experiences these illusions is not. It's separate from the deceiver, separate from the deceptions. And what is my "I"? It's outside the simulation. So Cartesian simulation doesn't explain all of reality. It leaves the deepest mystery, the "I", unsolved.
So simulation theory must be saying that it's my "I" itself, my very consciousness, the thing (whatever it is) that experiences my experience, that is also being simulated. And there, Bostrom's argument fails. The premise is that 40 years ago we had crude video games like Pong, and today we have hyper-realistic video games, and in the future they'll be indistinguishable from reality. I have no doubt that's true.
But Bostrom does not explain how consciousness itself will be implemented by a computer. We have no idea how to do that, nor any theory that says it's even possible. Turing machines are extremely constrained. They do one discrete thing at a time and can't solve the Halting problem. That's an example of a problem that we can conceive that no computer can solve. There's no evidence, only a belief on the part of some, that the universe is so constrained. Some people think Turing machines can be conscious, but nobody knows how.
Bostrom's premise fails. In the future we'll have super-duper realistic experiences to enjoy. But there's no evidence or even a convincing argument that subjective consciousness can be implemented. The argument is just wrong. And it's a bit disingenuous, because it begins with the Pong versus modern video games analogy, but sneakily ignores the problem of subjective experience. Descartes did not miss that point. He noted that even if the great computer in the sky is programming my reality, it can't program my subjective experience. I exist separately from the simulation.
Bostrom's analogy from Pong to modern video games is false. The electronic paddles in Pong did not have subjective consciousness; nor do the avatars in modern video games. In terms of the core mystery of subjective experience, no progress at all has been made in the past 40 years. And that destroys Bostrom's premise.
Please see my other response to this question here, which makes similar points.