F# hits a sweet spot that's hard to explain until you've used it: you get the conciseness of a scripting language but with a type system that catches errors before you run anything.
For automation specifically, the pipe operator (|>) makes data transformation chains read almost like pseudocode. Compare that to Bash where you're constantly context-switching between quoting rules and subshell behavior.
The thing that keeps me coming back is that F# scripts scale up. A quick .fsx file can grow into a proper project without rewriting anything — same language, same idioms. With Python you often hit a point where the "script" needs to become "real code" and there's a refactor involved.
The downside is real though: cold start on dotnet fsi is noticeable, and the ecosystem for quick one-liners isn't as mature as Python's. For anything touching data science or quick API calls, Python still wins on available libraries alone.
But for internal tooling, build scripts, and anything where correctness matters more than iteration speed — F# is underrated.
