Given that a bad Windows update now has the potential to literally render millions of systems inoperable without their users being able to do much about it, one would hope there are processes in place so no such trivial mistake by a single person could actually get as far as the production systems.

True, but if there was no such process, then you shouldn't fire the engineer who made the mistake, but instead consider punishing the person responsible for not setting up said process.

We set up processes not so that we can have an excuse for firing people who made errors, but so that those errors have limited consequences and could be safely used as learning opportunities.