Are you expected to run five Python type-checkers now?

The article argues that Python library maintainers should prioritize running multiple type-checkers (mypy, Pyright, Pyrefly, etc.) on their test suites and public APIs rather than extensively on internal source code. This approach is more valuable to users, whose experience depends on whether the library's public API works correctly across different type-checkers, while avoiding the code pollution from numerous type-ignore comments needed to satisfy all checkers' quirks simultaneously.