Is Python's weak typing a liability when generating code with AI?

I've been using Claude/Cursor heavily and noticed something: with TypeScript, the compiler catches interface mismatches immediately. With Python, I'm running code to find out it breaks.

This matters more with AI-generated code since I'm less familiar with what it wrote. Static typing gives me a safety net - I can see if the pieces fit before running anything.

Python dominates AI tooling, but is that despite the typing, not because of it? Or am I overvaluing compile-time checks?

(Yes, I know about type hints. Curious if anyone actually enforces them in AI-assisted workflows.)