Not really your place to judge is it? What if they are in the middle of a long running task and the machine shuts down? Maybe it takes all day to run that process intensive task? Boom updates reboot the machine and a whole day's processing was lost. It is the users' prerogative on how they want to use their machine (poor software design choices and all) and if they don't want to take an automatic update and reboot then so be it. Or what if some people like to wait a couple weeks and see if Microsoft accidentally breaks something major with an update... pretty sure that has happened before.

Lost a 3d print because of windows update.

Switched to Octopi after that.

If this happens every week and you haven't done anything about it, well...

It only has to happen once.

Besides which, a predictable failure is still a failure. By your logic, Win95 crashing several times a week wasn't a problem either.

Win95 crashing several times a week was a feature that would let you get away with not seeing this other bug: https://support.microsoft.com/en-us/kb/216641

This thread is predicated on the assertion that users lose 10 or even 20 hours per week to reboots. Not just once.

In my case, I started a simulation running overnight. The next day I found that Windows had rebooted to install a patch and killed my simulation. (I'm now using a Mac.)