Preventing AI development indefinitely via desperate politico-military struggle would just not work in the long term. Trying would be utter folly.

This reminds me of Charles Stross' "why even try - AI can't be stopped" (my paraphrase).

Surely if it buys a little extra time for a FAI singleton to succeed, desperate struggle to suppress other lines of dangerously-near-culmination AI research would seem incumbent. I guess this might be one of those scary (and sufficiently distant) things nobody wants to advertise.