OK, but this was a bug in an update of a kernel module that just happened to be deployed on Windows machines. How many OSs are there that can gracefully recover from an error in kernel space? If every machine that crashed had been running, say, Linux and the update had been coded equivalently, nothing would've changed.

Perhaps rather than an indictment on Windows, this is a call to re-evaluate microkernels, at least for critical systems and infrastructure.