// the find
sharkdp/fd
A simple, fast and user-friendly alternative to 'find'
fd is a faster, saner replacement for the Unix `find` command. It defaults to regex search, respects .gitignore, ignores hidden files, and uses parallelized traversal to hit ~13-23x faster than find on large directory trees. The target audience is developers who use find daily and are tired of its flag syntax.
Parallel directory traversal using the same ignore/regex crates as ripgrep — the benchmark numbers (854ms vs 19s on 4M files) are real and consistent. Smart case handling (case-insensitive until you type an uppercase character) is the right default and saves constant flag-passing. The -x/-X exec system with GNU Parallel-style placeholders ({}, {.}, {/}, {//}) is genuinely useful and handles parallel output without interleaving. Shell completion generation via --gen-completions rather than bundled static files is a good call — stays in sync with the binary automatically.
On Debian/Ubuntu the installed binary is fdfind, not fd, because of a name conflict with an unrelated package — you have to manually symlink it, which is a papercut that bites every new Linux user. The .gitignore-respecting defaults are great in a project directory but confusing when you just want to search /tmp or /etc and nothing shows up; the troubleshooting section exists but new users will still lose time to this. No built-in content search — you always end up piping to rg anyway, so the fd+rg combo is the real workflow but it's not a first-class interface. The exec parallelism has a documented race condition with nested paths (e.g. deleting all 'foo' dirs when foo/bar/foo exists), and the fix is just 'be careful' rather than a --no-parallel-exec-on-directories guard.