1. The comparison was to MySQL and Oracle storage using git diff format as an analogy, not git storage.

2. git storage does compress, and the compression is "diff-based" of sorts, but it is not based on commit history as one might naively expect.