I like that prediction, but git is actually a bit of a step back from older more traditional source control systems. Rewriting history is far more common with git to the point where teams have to make explicit policies on when that is allowed.
Even if git becomes common in science outside of code, I don't think it would overcome the social pressure to appear infallible.
> Rewriting history is far more common with git
How do you mean? It's impossible. You git push something, and then you leave an immutable trail that someone somewhere may have cloned so even if you were to force push or delete you can't erase it. Assuming it was of any importance and someone was watching (but if not of any importance, then no big deal to force push/delete).
Saying something is immutable because someone may have saved an old copy before you mutated your own copy sounds very fragile to my ears.
If you push to GitHub it won't delete it
Look in any field—epidemiology, computer science, physics, chemistry, medicine, et cetera. If the leading experts aren't using git yet, look for the ones that are. They will be the leading experts of tomorrow.