I started to ask this question in that thread:
Is an unpatched W7 system faster than a patched one?
...but opted not to. The tone of that thread did not lend itself to a calm discussion.
XP pre SP1 would run on a system with only 256MB. Try that with SP3+.
One XP patch caused a company that I work with some lost labor and files ($$$). Fortunately, it was a patch that could be uninstalled... but the files were gone for good. I got the kudos for finding the KB that caused the mess - no money mind you, just kudos :-)
To be fair to MS, the XP patch was not bad or flawed and the KB warned
about using it in an environment that had servers using SMB 1. But seriously, should every business have to have someone on staff that knows that SMB is? Much less what version their headless file server is running. And that KB came in via auto-updates. (If I recall correctly).
To be clear, I'm pro-updates (even if they slow a system down). But patches are not risk free. By definition, they change things. I would find it hard to calculate the odds of getting exploited vs. the odds of $damage$ due to a patch.
I lost many hours of labor just last week due to a W7 patch.... maybe I'm a jinx. :-(