I feel like Bostrom uses his first 2 conditions to constrain probabilities until the 3rd seems likely. Is this insightful or a trick? I would argue it's a lot closer to a trick, since conditions 1 and 2 are futures we hope for (that we survive and become technologically advanced)--so we tend to overestimate their actual probability.

Think of it this way: the total probability that we are in a simulation should be the sum of all future states in which our descendents create adequate simulations, divided by the total number of all possible future states--which is nearly infinite. Therefore the probability that we are in a simulation now is actually nearly zero.