To me, the two big things about ripgrep are (1) it's much more oriented towards the use case of searching a directory hierarchy full of code, eg by default it's recursive and respects .gitignore, it has a --type flag which knows various kinds of source files (eg --type c will only look at .c and .h files), etc, and (2) it's fast - uses all the lastest and greatest algorithms, and makes heavy use of parallelism. It's also just generally good and sensible.