Interesting. Some of the remote employees I manage and work with have some very mysterious commit times. But I don't think I have found any correlation with code quality and late night coding. The late night coders are sometimes more productive.
Are your remote employees writing kernel code which will go out to millions of machines with a high cost of failure coupled with a demonstrated history of poor quality?
If not, your definition of code quality may be different than what is being discussed here.
You've misunderstood what Linus is saying. It's not that the commit times are at odd hours, the point is that the commit times are too recent for the code to have had enough time to undergo much testing.