Bostrom's paper doesn't purport to show that we are probably in a simulation, but only the weaker claim that one of these things is true:

• Humanity is likely to fail to develop to superpowered "post-human" levels.
• Conditional on attaining superpowered post-human civilization, humanity is unlikely to run a lot of historical simulations.
• We are probably in a historical simulation.

(Bostrom puts it slightly differently; I think what I've written above is clearer and has fewer little holes.)

You will observe that this argument is more or less a triviality; Bostrom's contribution is thinking of making such an argument rather than filling in difficult steps in the reasoning once the argument is thought of.

I confess that my own response to this is indifference; I think there's a very good chance that the sort of computational superpowers needed to run a lot of faithful historical simulations will never be ours, and I don't see why a post-human civilization would bother to run a lot of simulations of their ancestors, so the most the argument can tell me is that it's not completely impossible that I might be in a simulation. Fair enough, but so what?

(Elaborating on that not-seeing-why: it's not very clear why our posthuman successors would bother running any ancestor-simulations, but to get "I'm probably in a simulation" out of Bostrom's argument what's necessary is either that the bit of my life I'm experiencing right now has been simulated not just once but many many times, or else that the posthumans are going to simulate not only their actual ancestors but many many people very like their ancestors in situations similar to their ancestors'. I see no reason to expect either of those.)