The most underappreciated risk in this piece isn't the megaprojects, it's the behavioral change that happens before they're built. If major powers simply believe AI may eventually enable nuclear primacy, they'll shorten decision timelines and lean harder on launch-on-warning before the capability gap is even decisive. The destabilization doesn't require the counterforce strike to succeed. It only requires both sides to start acting as if it might. That's a much shorter fuse than the article's mid-2030s technology timeline suggests.
The most underappreciated risk in this piece isn't the megaprojects, it's the behavioral change that happens before they're built. If major powers simply believe AI may eventually enable nuclear primacy, they'll shorten decision timelines and lean harder on launch-on-warning before the capability gap is even decisive. The destabilization doesn't require the counterforce strike to succeed. It only requires both sides to start acting as if it might. That's a much shorter fuse than the article's mid-2030s technology timeline suggests.
This post is obviously incorrect. Did you ask an AI for feedback on it?
Nuclear missile silos hidden under hundreds of meters of rock or soil cannot be reached by these strikes. Nice fear mongering though.