Many thanks for this, Rohin. Indeed, your understanding is correct. Here is my own screenshot of my private announcement on this matter.
This is far from the first time that Phil Torres references my work in a way that is set up to give the misleading impression that I share his anti-longtermism view. He and I had extensive communication about this in 2020, but he showed no sympathy for my complaints.
Is this slightly off? The factor that goes into the expected impact is the chance of being a non-sim (not the chance of being a sim), so for the argument to make sense, you might wish to replace "the chance of being a sim [...] is 1000x higher than..." by "the chance of being a non-sim is just 1/1000 of..."?