I dont have the full context, but I cant be bothered to read through all the back and forth (sorry membraan), so heres my two cents on accuracy et cetera.
Only one ways to get "pinpoint" accuracy (i.e. difference between impact point and intended impact point <<<<<< blast radius) - Have some sort of localisation method (meaning, providde a reference that can continuously help correct projectile guidance)
This can be laser illuminator on target for LGBs and Hellfires, can be a seeker on bomb itself that can identify the target, can be strong, jam resistant GNSS, can be loitering munitions using AI for target ID and then homing onto it, etc etc.
But plain INS guidance, even with best of ring laser gyros, will lead to a decent amount of drift that makes any "pinpoint" strikes unlikely.
(PS: Minor point, but hellfire has a range of 8 km, so that launch was likely much closer than 40km even if the non explosive warhead sheds some weight off the missile.)
Also note, jam resistant GPS modules exist, which will ignore ground based jammers/spoofers... Could just be that Pakistan failed to jam GPS guidance on our missiles?