You've misunderstood the essay. A central point is to not think about AI risk in terms of matrix/sci-fi scenarios, but in terms of how it might/is likely to function in our societies.

So if you discount the sci-fi scenarios (which is a separate argument), the argument that "we need to be the first to develop AGI" doesn't hold water. If, say, China gets there first, so what? It will affect their society however it'll affect it (likely entrenching authoritarianism if i had to guess). But this has no bearing on us (speaking as an american)