Well, as to me, times when it was okay to jump into production server and hack-hack-hack code there are mostly gone away, at least when we talk about revenue generating projects.
Just imagine – somebody would need to help or replace your current coder on this project, says he’s busy with another project now. With revision control system, specially if properly used (I mean if the commits get real comment describing what’s done, no giant pile commits like “yesterday I changed it all or at least it looks like all lines were changed”) with integration to issue tracking, so that it’s possible to open a ticket and found all related code changes, what impact do you imagine can this have?
As to me, it’s self-evident.
Forgotten to say and it’s exactly about forgetting – people tend to forget. Even author can forget what he did, why he changed those lines and so on. Revision control is of a great help here. 3 months later w/o quick going through history of commits you have more chances to loose even more time and make mistakes.
Git isn’t about collaboration only. It’s about development in general.